Abstract: In the situation of sample pollution, nulls would be formed in the expected direction, which results a great decrease in the signal-to-interference-noise ratio. And in the situation of limited snapshots, the given white noise exists in the form of color noise, so that there would be an increase in the distortion of adaptive beamforming. Based on these, an adaptive method is presented in this paper by analyzing the eigenvalues of sampling covariance matrix. This method is equivalent to interfering the sampling covariance matrix in order to improve the performance and robustness of beamforming. Finally, performance analysis and computer simulation demonstrate the feasibility and effectiveness of this method.