Because uncertainty and complexity of the athlete’s injury risk factors, the general risk method can’t reflect the risk of athletes. Based on Bayesian probability estimation method, this paper proposes a new evaluation system of athlete injury risk. Combined with the VB programming and visualization display function, it realizes visual evaluation function of athlete injury risk, and the injury level of athletes is divided into three warning levels. Finally through risk assessment system, we get the effect curve of athletes training number and training time on the injury. From the results we can conclude that the training time has a certain randomness injury to the athletes, and training times are decisive factors to some injury of athletes. It provides technical reference and data support for the risk assessment of injury.