Abstract: In order to achieve simply and effectively the ideal edge of a medical image and thus perform an effective medical diagnose, a fast edge detection method based on the combination of fuzzy subsets is developed. We partition an image into two portions:the edge portion and the non-edge portion. Non-edge portion, consisting of the objects and its background, is removed from an image; the remainder is image's edge. Considering the fuzziness of image's edge, some fuzzy operations can be carried out. We firstly partition the gray-level histogram into several sub-regions and operate corresponding fuzzy subsets. Finally, image's edge is obtained. Experiment results show as compared with Sobel operator or Prewitt operator, that the described method is more excellent.