In recent years, the athlete mental health problems have become more and more prominent. Compared with high school, college athlete students are in a more complex environment and face more diversified difficulties, such as the constant emergence of problems in academics, interpersonal interactions, emotional life, employment, etc., which make college athletes overwhelmed, and the accumulation of pressures from many sides leads to a variety of psychological problems. Therefore, it is important to pay attention to the individual athlete mental health problems, and to detect and guide the development of athletes' mental health in time, which is crucial for the management of athlete students in colleges and universities. The purpose of this paper is to use data mining principles and methods to explore the factors leading to psychological problems of college athletes, so as to carry out psychological intervention purposefully. In this paper, we use data mining technology to set up a questionnaire with 90 questions reflecting mental health, and take SCL-90 Symptom Self-assessment Scale as the measurement standard to collect data, establish a data set, analyse the data using improved Apriori algorithm, and establish a decision model of obsessive-compulsive symptoms using decision tree ID3 algorithm, which provides a theoretical basis for the evaluation of the athletes mental health status of college athletes and decision-making of health checkups. It provides a theoretical basis for the evaluation of the mental health status of athletes and the decision-making of health examination.