Abstract: The sensor scheduling problem for a target-tracking sensor network is studied.The scheduling objective is designed considering tracking accuracy of each target and the overall tracking performance from the systematic viewpoint.The extended Kalman filter is adopted for target tracking and accuracy calculating.A mixed integral model is founded for this scheduling problem，the scale of which is too large to solve using traditional optimal methods.Thus a new genetic algorithm based on a local solution space jumping mechanism is designed to solve the model.Experimental results show that the presented algorithm performs well on this problem.