The psychological problems of college students under the rapid development of big data on the Internet have aroused widespread concern in society, and mental health affects the normal learning life of college students. This paper analyzes the psychological status quo of college students using the internet, proposes a mining process for mental health education data, and performs data preprocessing. The application of the Apriori algorithm in data mining was analyzed, and it was found that the algorithm has problems such as too long matching times and too many item sets. Improvements were made on this basis, and the improved Apriori algorithm was applied to the analysis of college students’ psychological problems. It was found that interpersonal tension and sensitivity, depression, and anxiety were the most common psychological problems. In addition, in the teaching mode comparison experiment, the Internet+ mental health education program based on association rules had a good influence effect on students. Students’ physical and emotional feelings exceeded 4.00 points, and their satisfaction with the course reached 4.65 points, so the internet+mental health education course based on association rules has some promotion value.