To smooth out speckle noise and preserve edge information in noisy images, a novel algorithm called speckle reduction by adaptive window anisotropic diffusion is proposed. Based on the image's structure, the proposed method can adaptively modulate its window size and orientation. This adaptive window method can calculate the instantaneous coefficient of variation for the edge area with more accuracy than a fixed window method. Furthermore, a novel method for finding the diffusion coefficient is presented. The experimental results show that compared with existing methods, the performance of the proposed method is satisfactory in terms of both speckle suppression and preservation of image details.