Abstract: The Earth's magnetic field is well modeled and is a function of the satellite's position,therefore,autonomous Low-Earth-Orbit(LEO) satellite navigation can be achieved based on the measurement data of the onboard three axis magnetometer(TAM).High precision satellite motion model was presented firstly,then autonomous navigation algorithm by using measurement information of star sensor and TAM was presented.The observation equation was established by using two observation models and integrating measurement vectors of star sensor and TAM based on Federal Kalman Filter.The algorithm convergence and numerical simulation precision were analyzed in detail.The results of computer simulation show that the proposed system can improve the navigation precision,robustness and reliability.