The process of English learning is closely related to the process of intercultural communication. This paper addresses the problem of quantifying the intercultural communication ability of college students, designs the dimensions of intercultural competence, establishes the factor indicators of intercultural competence, and constructs a fuzzy comprehensive evaluation method for the assessment of intercultural competence. A regression model is created to investigate the correlation between achievement and English proficiency. The nonparametric model of principal component analysis is used to optimize the problem of the stratification effect of general local linear regression, and the principal component-based local linear regression method (PC-LLR) is proposed. The second-year undergraduate students of different majors in X university were selected as the research subjects, and an overall assessment of intercultural communication competence was made to explore the differences among different student groups. Based on the correlation analysis between intercultural competence and English proficiency, PC-LLR regression analysis was launched. The Beta value interval of the regression was [0.0535, 5.0292], all of which were positive, and the four variables of knowledge, attitude, skill, and awareness had a positive predictive effect on the dependent variable. Intercultural knowledge has a relatively high Beta value, which makes it a strong predictor of English proficiency.