Abstract: In this paper, a noise cancellation method using microphone array in the presence of interference noise and background noise is studied, where the signal directivity is utilized. Beamforming captures the speech in the specified direction and suppresses the noise in other directions. After beamforming, however, there will still be some residual interference noise in the enhanced speech. To overcome this problem, this paper proposes a generalized sidelobe cancellation beamforming method that uses the signal power spectral density ratio to further achieve the suppression of background noise and interference noise. In addition, this paper utilizes deep neural network to improve the logarithmic spectral amplitude estimation by training the masking value under the multi-objective function. After that, the post-filtering is used to eliminate the residual interference noise. Finally, different baseline methods are compared in the experiments, and the effectiveness of the proposed algorithm is also verified.