Pseudoscientific language comparison is a form of pseudo-scholarship that aims at establishing historical associations between languages by naïve postulations of similarities between them.

While comparative linguistics also studies the historical relationships of languages, linguistic comparisons are deemed pseudoscientific when they are not based on the established practices. Pseudoscientific language comparison is usually performed by people with little or no specialization in the field of comparative linguistics. It is a widespread type of linguistic pseudoscience.

The most common method applied in pseudoscientific language comparisons is to search two or more languages for words that seem similar in their sound and meaning. While such similarities often seem convincing to laypeople, linguistic scientists see this kind of comparison as unreliable for two primary reasons. First, the method applied is not well-defined: the criterion of similarity is subjective and thus not subject to verification or falsification, which is contrary to the principles of the scientific method. Second, the large size of all languages' vocabulary makes it easy to find coincidentally similar words between languages.

Because of its unreliability, the method of searching for isolated similarities is rejected by nearly all comparative linguists (however, see mass comparison for a controversial method that operates by similarity). Instead of noting isolated similarities, comparative linguists use a technique called the comparative method to search for regular (i.e. recurring) correspondences between the languages' phonology, grammar and core vocabulary in order to test hypotheses of relatedness.

Certain types of languages seem to attract much more attention in pseudoscientific comparisons than others. These include languages of ancient civilizations such as Egyptian, Etruscan or Sumerian; language isolates or near-isolates such as Basque, Japanese and Ainu; and languages that are unrelated to their geographical neighbors such as Hungarian.