Aiming at the robustness of the data fusion algorithm of GPS/SINS combined navigation on Unm anned Aerial Vehicle(UAV) and the requirements of real-time solution, the initial sampling value is selecte d based on the unscented Kalman filter to solve the se nsitive problem and 2n+ 1 Sigma sampling points are required for the solution, which will make the solutio n efficiency low. The Iterative Spherical Simplex Unt racked Kalman Filter (ISSUKF) is proposed, and the supersphere distribution sampling strategy is adopte d, and the sampling point Sigma is reduced to n+2 by changing the Sigma point sampling strategy, and the state estimation obtained by the measurement equati on is used to iteratively update to reduce the influenc e of the initial state value on the system, thereby redu cing the calculation amount and improving the stabili ty and accuracy of the algorithm. The simulation resu Its show that the ISSUKF algorithm has the character istics of high accuracy, high computational efficiency and good robustness.