Abstract: Global navigation satellite system is widely used in military, communications and other fields, but the navigation signal is vulnerable to electromagnetic interference in space, resulting in serious performance degradation. Based on the method of signal separation, the problem of interference mitigation in the case of missing samples is studied, a sparse observation model by decomposing the amplitude modulated signal of interference signal is constructed, sparse Bayesian learning is used to reconstruct the interference signal, the reconstructed interference signal from the received signal is further removed, and the navigation signal after interference mitigation is obtained. The simulation results show that the proposed method can effectively suppress the interference in the navigation signal and improve the quality of the navigation signal. The proposed method still has good interference mitigation performance in the case of missing samples.