Abstract: To investigate the autonomous navigation of probe orbiting the Mars orbit X-ray pulsars, different real-time adjustment methods are applied in this paper, including the extended Kalman filter (EKF), adaptive extended Kalman filter (AEKF) and robust adaptive extended Kalman filter (RAEKF). Firstly, the observation data, i.e., time of arrival, is simulated according to the principles of pulsars navigation. Secondly, extended Kalman filter (EKF), adaptive extended Kalman filter (AEKF) and robust adaptive extended Kalman filter (RAEKF) are used to get the position and velocity of the probe. Thirdly, the adjustment results are compared with the orbit generated by STK, the results show AEKF and RAEKF performs better than EKF, the result of AEKF is better than the km level of VLBI and can reach the precision level of triple pass Doppler, leveling at 94.1 m in X-axis, 19.5 m in Y-axis and 22.3 m in Z-axis.