This paper focuses on the image segmentation, which is one of the key problems in medical image processing. A medical image segmentation method is proposed based on the combination of fast marching method and watershed transformation. First, the original image is smoothed by using nonlinear diffusion filter. Then the smoothed image is over-segmented by the watershed algorithm. Finally, the image is segmented automatically by using the modified fast marching method. Moreover, the speed function is defined based on the staristiical similarity degree of the regions.Experimental results show that the algorithm can obtanin segmentation result of medical image fast and accurately.