Abstract: Time difference of arrival (TDOA) and frequency difference of arrival(FDOA)estimation is the basis of multi-satellites passive location, and the estimation accuracy determines the performance of passive location system. In order to solve the problems of high computational complexity, poor real-time performance and limited estimation accuracy of current TDOA/FDOA estimation methods, a high precision and fast estimation method for TDOA/FDOA based on multi-stage search is proposed. In the proposed method, the original data received by the two satellites is firstly segmented based on the zoom fast Fourier transform to decrease the search range of FDOA, which can further reduce the computational complexity. Secondly, the segmented data is extended forward and backward to avoid the correlation loss caused by segmented processing. On this basis, the cross-ambiguity function of the two signals is calculated, and the coarse estimation of TDOA/FDOA is obtained through spectral peak search. Finally, based on the coarse estimation of TDOA/FDOA, the fine estimation of TDOA/FDOA is obtained by using the signal correlation spectrum thinning method. The computer simulation results show that the proposed method can effectively reduce the computational complexity while improving the accuracy of parameter estimation. Therefore, the proposed method can provide technical support for subsequent engineering applications of multi-satellite passive location.