Abstract: In order to solve the problem that various errors are accumulated over time and the long-term accuracy of the system is poor, a dual-track fusion navigation system based on MEMS sensor is proposed to solve the problem of autonomous pedestrian inertial navigation.The zero-speed state is detected according to the zero-velocity detection method, and the zero-velocity updates (ZUPT) correction is carried out by using the Kalman filter.Finally, the maximum separation distance of the pedestrian left and right foot is taken as the constraint condition to realize the double trajectory Fusion.Experiments show that the overall navigation performance is improved by using the proposed navigation system scheme, and the position error is reduced by 65% compared with when not being integrated.