To spot a confusingly similar trademark in Nepal, compare two marks the way a hurried buyer would: how they look, how they sound, what they mean, and whether they sit on the same or related goods. The Department of Industry (DoI) refuses and cancels marks on those grounds, so check before you file.
Key takeaways
- The Patent, Design and Trade Mark Act 1965 never uses the words "similar" or "confusing"; the test comes from Sec. 18(1) and the courts.
- The Supreme Court calls copied letters with a small change or addition deceptive similarity (Decision No. 10561).
- IP Sewa's archive holds 1,573 same-class pairs of published names that differ by a single letter.
- Say names aloud: b and v, sh and s, and doubled vowels hide sound-alikes that a spelling search misses.
- A published look-alike of your mark can be opposed within 90 days of its Bulletin publication.
- Our free search is deliberately not fuzzy, so type each variant yourself.
What does "confusingly similar" mean under Nepal's law?
The phrase is a practitioners' label, not statutory wording. Counted over the English consolidation of the Patent, Design and Trade Mark Act 1965, the words "similar", "confusing" and "deceptive" each appear zero times. The working test sits in the proviso to Sec. 18(1).
Under that proviso, the DoI will not register a mark that may hurt the prestige of a person or institution, harm public morality or the national interest, or undermine "the reputation of the trade-mark of any other person". Nor will it register a mark "already registered in the name of another person".
Sec. 18(3) then lets the DoI cancel a registered mark where any of those circumstances exists, after giving its owner a chance to show cause. A look-alike can therefore fail at examination or be removed later. You can read the Act through the Nepal Law Commission.
What did the Supreme Court say about deceptive similarity?
The leading case is Kansai Nerolac Paints Ltd. v. Rukmani Chemical Industries Pvt. Ltd., Decision No. 10561, reported in NKP 2077 and decided by a joint bench on 2076/11/07 BS. Its headnote defines deceptive similarity.
A mark is deceptively similar when it keeps the same letters, copies some letters in the middle or at the end, or adds a little, so that the two names sound almost alike and look the same at a glance.
A second headnote point matters to brand owners: where a foreign company's mark is registered with bad intent and in bad faith, no limitation period stops its cancellation. Our guide to how Nepali courts decide trademark similarity covers the wider case law.
What does the register show about the NEROLAC dispute?
The Court reversed the DoI and the High Court Patan and left the local company exclusive rights only in the letters "RCI". It ordered ten marks published in Bulletin Year 6, Number 5 (2069/05/14 BS) to be registered in the name of Kansai Nerolac Paints of Mumbai.
IP Sewa's bulletin archive shows both sides on the page. RCI NEROLAC entries filed by the local company were published in August 2011 and again in 2012, and our copy of Volume 6, Number 5, published on 30 August 2012, holds exactly ten Kansai Nerolac Paints Limited entries in Class 2, the paints class.
How close do published names get in Nepal?
Closer than most founders expect. Across IP Sewa's 71,500+ published marks, we found 1,573 pairs of names in the same NICE class that differ by exactly one letter. Class 30, the staple-foods class, has the most, with 261 pairs.
The basis: Latin-script names of five or more letters, compared without spaces or punctuation, leaving out entries that only describe a device, counted on 17 October 2026.
Some pairs are one owner's own variants, filed on purpose. Many belong to different businesses, like those below.
| What changed | Published pair, different owners | Class | Years published |
|---|---|---|---|
| b and v swapped | SARBOTTAM / SARVOTTAM | 3 (soaps, cosmetics) | 2025 / 2012 |
| h dropped | SHIKHAR / SIKHAR | 9 (electrical goods) | 2014 / 2006 |
| vowel doubled | SHAKTIMAAN / SHAKTIMAN | 7 (machines) | 2018 / 2026 |
| British or US spelling | COLOUR MAGIC / COLOR MAGIC | 2 (paints) | 2012 / 2026 |
| plural s added | VEDANTA HERBAL / VEDANTA HERBALS | 3 (soaps, cosmetics) | 2014 / 2026 |
These pairs prove nothing about who was first, whether either mark was opposed or registered, or whether anyone copied. Close names can coexist lawfully, for example where the goods differ within a class. They show how little can separate names the DoI has already printed.
Which four checks reveal a look-alike?
Run four checks on every close result. Sight: compare spelling, letter shapes and the logo. Sound: say both names aloud, in English and in Nepali. Meaning: ask whether a translation or a picture carries the same idea. Goods: check whether the products reach the same buyers.
The courts look at the overall impression, so a shared dominant word can outweigh a different logo, and a near-identical picture can outweigh different words. Our full guide to the trademark similarity test covers each factor.
Why does sound matter so much for Nepali brand names?
Many Nepali brand names are transliterated, and there is no single standard for spelling Nepali sounds in Roman letters. B and v, sh and s, a and aa, and an added h all give two spellings of one spoken word, and a buyer ordering by phone hears only one.
The register shows the effect. Among published names, 71 contain GORKHA and 37 contain GURKHA, and three spell HIMALAYA as HIMALYA. Our search links five hand-checked groups like these (Himalaya, Gurkha, Everest, Annapurna and Sagarmatha), so typing one spelling also finds the other.
Outside those groups, matching is deliberately not fuzzy: a search for SARVOTTAM will not find SARBOTTAM. Search the Devanagari form too, because a Roman mark and its Devanagari twin can be the same brand to a buyer.
Does adding "Nepal" or a second word make a mark different?
Rarely. The Supreme Court's own definition covers a mark made by adding a little, so a copied core word with "Nepal", "Gold" or "Plus" attached can still be deceptively similar. Buyers remember the distinctive part and skip the descriptive tail.
Descriptive words are also crowded. In our archive, 662 published names contain NEPAL and 323 contain HIMALAYA, so neither word distinguishes a brand on its own. Judge your mark by its strongest word, and search that word alone.
Do the goods have to be the same for a conflict?
Not always. Nepal uses the NICE classification of 45 classes, the international list kept by WIPO, and Sec. 18A(2) needs a separate application for each class, but confusion follows buyers rather than class numbers. Paint and a paint shop, or tea and a café, sit in different classes yet meet the same customers.
The reputation ground in Sec. 18(1) is not tied to a class in its wording, so a well-known name can matter outside its own class. Start with the NICE class finder, then read choosing the right trademark class before deciding which neighbouring classes to check.
How do you spot a confusingly similar trademark, step by step?
Work through six steps for every shortlisted name, before paying for packaging.
- List the variants. Write your name, its likely misspellings, b/v and sh/s swaps, plurals, joined and split forms, and the Devanagari version.
- Search each one. Run every variant through IP Sewa's free trademark search and the trademark conflict checker, which ignore spaces and word order.
- Filter by class. Use search by NICE class to read every close name in your class and its neighbours.
- Check the owner. The owner search shows whether one company already holds several variants of a name.
- Compare the logos. Read the printed label of each close result; our logo similarity search, on a paid pass, screens artwork.
- Record your reasoning. Note each close result, its class and bulletin, and why buyers would or would not confuse it.
Careful owners close the gaps themselves. One appliance company filed both NIBOR and NIVOR in Classes 9 and 11, and all four entries appeared in the same 2024 issue and closing the b/v gap.
What can a free search not see?
IP Sewa's search covers marks published in DoI bulletins since Volume 1, Number 1 in January 2006, more than 120 issues. It shows what was printed, not current registration status, and it cannot see applications still waiting for publication.
That wait is real. Across the 9,652 marks in the 12 latest issues, the median gap from filing to publication was about 3.5 months, and about one mark in six waited over two years. An unopposed application typically takes about 12–14 months from filing to certificate, and 6–8 months only in the smoothest cases.
Those unpublished filings matter, because the first-to-file principle generally applies in Nepal and an earlier complete application can still block you. For a written view of the risk, our trademark search and clearance service checks the close results with you.
What if you spot a look-alike of your own mark in the Bulletin?
Act inside the opposition window. A published trademark can be opposed within 90 days of its publication in the Industrial Property Bulletin, counting the publication day as day 1, so the last day is the publication date plus 89 days. Rely on the deadline the DoI states, and file before it.
For the newest issue in our archive, Volume 20, Number 8, published on 16 October 2026, that last day is 13 January 2027. Count from the issue date, not the day you notice it.
New issues appear on the DoI's Bulletin page, and our trademark opposition guide covers the grounds, evidence and steps.
Gather your registration certificate, a power of attorney if an agent files for you, the Bulletin entry and dated sales records. Our opposition and enforcement team can review a conflicting entry and file within the window.
Can you still act after the 90 days have passed?
Yes, through a harder route. Sec. 18(3) lets the DoI cancel a registered mark where a Sec. 18(1) ground exists, after its owner has had a chance to show cause.
Kansai Nerolac adds that no limitation period bars cancelling a mark registered in bad faith. Proving bad faith takes evidence, so treat this as a fallback.
Found a close name: what should you decide?
Two questions settle most cases. Do the goods meet the same buyers? If they do, could a buyer mix the names up by sight, sound or meaning? A yes to both means renaming now costs far less than a dispute later.
Common mistakes: why do owners miss look-alikes?
Most misses come from searching too narrowly or reacting too late.
- Searching one spelling, so a b/v or sh/s twin slips past.
- Comparing logos only, while the spoken names are nearly identical.
- Trusting an added word such as "Nepal", "Gold" or "Plus" to create distance.
- Checking one class when buyers meet the goods in a neighbouring one.
- Reading a published pair as proof that close names are always safe.
- Waiting past day 90 to react to a Bulletin entry.
What does a near-miss look like for a new brand?
Here is an illustration, not a real case. A Pokhara founder plans to sell packaged tea in Class 30 as "Venoka". Her search for VENOKA finds nothing close, and she is ready to order packaging.
Then she types BENOKA, the b/v twin, and finds an earlier published mark for tea from another business. Same class, same buyers, almost the same sound: she renames before printing anything, which costs far less than a rebrand after an opposition.
In short: how do you spot a confusingly similar trademark?
Compare sight, sound, meaning and goods; search every spelling, not just yours; and read each new Bulletin for look-alikes of your mark. The register already holds many names one letter apart, so a careful search before filing is the cheapest protection a brand gets.
A registered mark lasts seven years from registration and is renewable any number of times, so repeat the check whenever you add products or classes.
People also search for
- How to search for a trademark in Nepal
- Why trademark applications are rejected in Nepal
- How Nepali courts decide trademark similarity
- Trademark opposition in Nepal
Ready to clear a name? Start with our trademark registration service, run your variants through the trademark search, or contact IP Sewa with the close results you found.














Comments (0)
Leave a comment
Replying to — cancel