In the context of globalization, the number of international students in vocational colleges is increasing year by year, and their demand for Chinese learning is increasing. As the most populous language in the world, the mastery of Chinese language directly affects the competitiveness of international students in the professional field. However, due to language and cultural differences, international students often face many challenges in learning Chinese language, such as difficult pronunciation difficulties, complex grammar, and lack of cultural understanding. Fuzzy clustering algorithm, as a data mining method capable of dealing with uncertainty and ambiguity, provides new possibilities to solve these problems. Through this algorithm, students’ learning characteristics can be identified more accurately, and personalized teaching plans can be formulated, so as to improve the teaching efficiency and learning results.MATLAB simulation shows that under the condition of certain evaluation criteria, the fuzzy clustering algorithm is better than the existing evaluation methods for evaluating the quality of training of international students in China.