The mental health problems of college students are becoming increasingly prominent. It is of great significance to detect the tendency of psychological problems as early as possible and reduce the occurrence of psychological crisis problems in order to promote the healthy development of students. Psychological health, as an important factor in the growth process of college students, is closely related to emotional changes. Social media is an important channel for college students to express their emotional attitudes, and the emotional information contained in their texts also provides data support for assessing college students’ mental health. In order to analyze the mental health status of college students more intelligently and accurately from their social texts, this study constructed a text sentiment analysis model for college students’ mental health, using machine learning and sentiment lexicon methods to visually display the trend of users’ emotional states from two aspects: emotional polarity and emotional intensity. This study will provide a reference for identifying the mental health status and warning of psychological crises among college students.