An effective de-noising method in edge protecting was proposed to overcome the limitation of the current image de-nosing methods, which combines edge detection with neighborhood weighed window. This algorithm decomposes noisy images using stationary wavelet transform to keep phase invariance. Then it detects the edges of low frequency sub-band and high frequency ones, and gets the approximate information of the edge of original image by fusion the results of edge detection. Based on intra-scale dependency and directivity of wavelet coefficients,the method filters different sub-bands coefficients using corresponding forms filtering templates with weighed window, except the information of edge. A better restoration of image is demonstrated in the results of experiments, with detail of images kept as well as image noises decreasing.