Kalman filter technique is one of the theoretical foundations of target tracking. In engineering applications, nonlinear filtering methods are mostly used to track targets. The Unscented transform is introduced, and the Unscented Kalman filter (UKF) is used to simulate the target tracking, which effectively overcomes the shortcomings of the traditional extended Kalman filter algorithm (EKF). The second-order approximation of the best estimate is given, which improves the accuracy and stability of target tracking. The simulation results show that the algorithm has good effects on the position and speed tracking of the target. For target maneuvering, the multi-model filtering algorithm can be used to achieve accurate tracking of target maneuver.