In order to improve the tracking performance of radar, a method based on the BP neural network model is given, aiming at the problem that the tracking task priority designing of traditional method cannot ensure radar always tracks important targets under multi-target trace. The method utilizes the target parameters and the performance parameter of radar to comprehensively judge the priority of the target tracking task directly, and adopts BP neural network to approximate the function relation between the parameters and tracking task priority, to realize rapid real-time design of tracking task priority. Finally, comparing with the method based on the fuzzy logic, the method in this paper is simulated. The analysis result proves the effectiveness and superiority of the method.