Medical ultrasound images suffer from speckle noise and low contrast. A novel anisotropic diffusion filtering, which is a modification of the conventional nolinear coherent diffusion filter, is proposed. The proposed filter suppresses the speckle noise in a homogeneous region using diffusion, and enhances the contrast by enhancing edge boundaries in a structure region. The input image is decomposed by Laplacian pyramid, and then, the proposed filtering is done during reconstructing the Gaussian layers to produce the filtered images. It has been shown by experiments that the proposed method enhances the edge and contrast while reducing noise.