Firstly,a digital shearlet transform is designed by Haar wavelet filter.And then,correlation of shearlet coefficients is statistically analyzed,and an adaptive threshold shrinkage denoising algorithm is proposed.Lastly,the experiments is carried out to make comparasion with wavelet and shearlet denoising method using peak signal to noise rate （PSNR）.The experimental results show the proposed algorithm can achieve the high PSNR value.Meanwhile,the textures and edges can be preserved better,and the denoised image is satisfied with visual quality.