The SAR image navigation system is highly nonlinear when it takes long time to output measurement in the SAR integrated navigation system (SAR/INS), and when the inertial navigation error cannot be modified in time. Using conventional Kalman filtering to deal with the nonlinear system, the filter accuracy is affected and even becomes divergent. In order to improve the integrated navigation precision, based on SAR/INS coupled navigation system and the unscented Kalman filtering (UKF), we propose a UKF algorithm for unequal interval problems, and solve problems of information fusion in SAR/INS. Performances of conventional Kalman filtering and the UKF algorithm are studied in simulation analyses of covariance. The result
demonstrates that the method proposed in this paper produces better filtering precision.