The multipath estimation of global navigation satellite system (GNSS) is actually a state estimation problem. Extended Kalman Alter (EKF) is a good estimator for state estimation in linear systems. However, the estimation results of EKF fluctuate obviously around actual values. To solve the problem, a sliding average EKF is proposed. In the sliding average EKF, EKF filtering results are smoothed by sliding average to decrease the fluctuation amplitude of the filtering results. The simulation results show that the estimation results of the proposed algorithm have smaller fluctuation amplitude compared with EKF.