A sensor resources scheduling algorithm is proposed to improve the track accuracy and system work effectiveness of multi-target tracking problem based on warning system of LEO. Firstly, the target tracking model is constructed. Then, with two contradictory objectives, the PCRLB change and switch rate of satellite during the scheduling period, a mixed integral programming model is founded. Based on the model, a mixed genetic algorithm and simulated annealing is proposed to solve the sensor scheduling problem. The algorithm improves the seeking ability and convergence speed. Finally, simulations demonstrate the correctness of the model and the validity of GA-SA algorithm.