As one of the best denoising method, the Block-matching and 3-D filtering (BM3D) has a very important role in image denoising. But the research on basic of BM3D did not make much progress. In this paper, we propose a method to improve the algorithm by the complex wavelet. The real wavelet transform has two problems, shift variance and lack of directionality. The complex wavelet transform can solve the problem of real wavelet transform which used in BM3D. By replacing the real wavelet transform with the complex wavelet transform, we obtained improved denoising results. Experimental results demonstrate that the improved algorithm achieves a better result than the original method in terms of both image denoising and peak signal to noise ratio (PSNR) for test image.