With the development of information technology and the popularization of artificial intelligence technology, the basketball field is gradually using computer means to solve the problems it faces. At present, sports athletes cannot effectively select suitable basketball sports according to their own characteristics and preferences, resulting in a decline in athletes' enthusiasm. In response to this problem, this research proposes a basketball risk prediction model from a personalized perspective. First, analyze the shortcomings and limitations of traditional collaborative filtering methods; secondly, improve the traditional collaborative filtering algorithm and establish a risk prediction model. Athletes' risk degree and item attribute scores are used to predict weighted risk; finally, the actual data of sports athletes are used for model analysis and verification, and the advantages and disadvantages of the model proposed in this study and the traditional model are compared. The experimental results show that the improved collaborative filtering technology has good accuracy and efficiency in predicting the risk of basketball players.