The invention discloses an image de-noising method based on a two-dimension indivisible DFT modulation filter bank. The method comprises: using the two-dimension indivisible DFT modulation filter bank for two consecutive times to perform local Wiener filtering; through the matching with the frequency directions of the sub-band filter bank to determine two oval windows so as to further work out the image's sub-band energy distribution; and finally obtaining the sub-band coefficient of the de-noised image. According to the invention, through the comparison of the calculated Peak Signal to Noise Ratio (PSNR) of a normally used testing image with the PSNR of the de-noised image based on the two-dimension indivisible DFT modulation filter bank, the competitive PSNR can be obtained and the de-noised image possesses rich and non-uniform texture details; and therefore, an image with better visual effect can be obtained.