This paper mainly studied the problem of the noise suppression for images corrupted by different kinds of noises. A weighted average filter based on the fuzzy theory is presented. The algorithm first looks on the pixels in the filter window as elements of a fuzzy set

and then optimizes the membership of each element of the fuzzy set by iteration ways. In the end

the filter gets its weights of the pixels in the window from the memberships of the fuzzy set. Computer simulations show that the presented filter is not only better than average filter in the field of suppression of Gaussian noise

but also good at the suppression of impulse noise. So it has good performance to image corrupted by mixture noise.