Abstract: The problem of automatic selection of window size is investigated for image thresholding by index of fuzziness. Under a predetermined membership function and given the class numbers of image pixels, an adaptive window size selection method for image thresholding by index of fuzziness is put forward. At the same time, images which are difficult to be segmented using thredholding by index of fuzziness because of histograms possessing one mode or two modes with great disparity in between are studied. A histogram transformation method is presented, and through the transformed histogram, images can he segmented effectively using the adaptive thresholding by index of fuzziness. Finally, examples show that the method put forward in this paper is simple, effective, and robust.