Abstract: For the blind source separation algorithm based on the maximum signal-to-noise ratio, the signal and interference cannot be effectively separated in the environment of poor signal stem noise. In this paper, an improved blind source separation algorithm based on the maximum signal-to-noise ratio is proposed, which uses the observed signal after the denoising of the wavelet threshold as a new observation signal, reconstructs the maximum signal-to-noise ratio objective function, and realizes the interference separation under high signal-to-noise ratio conditions. Through computer simulation, the separation of diffuse interference under high signal-to-noise ratio conditions is realized in the frequency diverse array-multiple input multiple output radar system, and the results show that compared with the original algorithm, the improved algorithm can achieve the separation of target and interference signal well.