Can my URLs use non-English words? John Mueller answers this question on Google's SEO Snippets video series.
For sites that target users outside of English-speaking regions, it’s sometimes unclear if they can really use their own language for URLs, and if so, what about non-English characters?
Google search uses URLs primarily as a way to address a piece of content. We use URLs to crawl a page, which is when Googlebot goes to check the page and to use the pages content for our search results.
As long as URLs are valid and unique, that’s fine.
For domain names and top-level domains non-Latin characters are represented with Unicode encoding. This can look a little bit weird at first. For example, if you take Mueller, my last name, with the dots on the U, that would be represented slightly differently as a domain name. For browsers and for Google search, both versions of the domain name are equivalent; we treat them as one and the same. The rest of the URL can use unicode utf-8 encoding for non-Latin characters. You can use either the escape version or the unicode version within your website; they’re also equivalent to Google.
Regardless of what you place within your URLs, make it easy for folks to link to your pages. For example, avoid using spaces, commas and other special characters in the URL. They work for Google, but they make linking a little bit harder. Use dashes to separate words in your URLs. Sme prefer using underscores; that’s fine, too. Dashes are usually a little bit easier to recognize. And if your site is available in multiple languages, use the appropriate language in URLs for content in that language.
So to sum it up, yes, non-English words and URLs are fine, [and] we recommend using them for non-English websites.