Abstract: To improve the calculation speed of high accuracy positioning estimation of global positioning system（GPS）, four nonlinear filtering algorithms were described and compared in theory, which are unscented Kalman filter (UKF) and its modified algorithm, spherical unscented Kalman filter (SUKF) and its modified version. Then the modified SUKF was proposed as an optimal estimation algorithm in the standalone GPS positioning. The experimental results show that the modified SUKF achieves high positioning accuracy and improves the calculation speed significantly; it can effectively solve the realtime problem with high positioning accuracy in GPS software receiver.