Abstract: The single snapshot DOA estimation method can solve the rank loss problem faced by traditional methods when short-term burst signals and coherent sources occur. In this paper, the covariance matrix is estimated by reiterative super-resolution technique, and the DOA is estimated by root MUSIC algorithm. This method does not need to search spectrum peaks, and can realize single snapshot DOA estimation without losing the array aperture. In this paper, the reiterative super-resolution technique based on the principle of minimum variance distortionless response is derived. The effects of spatial angle partition, the number of iterations and signal-to-noise ratio on the probability of separation and estimation accuracy are simulated and analyzed. The effectiveness of this method is verified by comparison with existing methods.