In order to solve the problem about filtering divergence of carrier under high dynamic motion, put forward strong tracking uncented Kalman filter (STUKF) algorithm. Analyze the combined navigation model and classic uncented Kalman filter (UKF) algorithm, use the STUKF algorithm in SINS/GNSS combined navigation system, and compared it with classic UKF algorithm and attenuation memory UKF algorithm. The analysis results show that the strong tracking UKF algorithm has good performance, can significantly shorten the filter time, reduce the speed error and position error, so it could improve the accuracy and stability of integrated navigation.