The task of image segmentation is a crucial step in image processing which is an important application of computer vision. The algorithm presented in this article is composed of two steps, which are the morph-based preprocessing and the ameliorated watershed method. In the first step, the foreground is simplified and the background is kept as the same as possible, and then the ameliorated filed-decision watershed method is used to achieve the segmentation of images in the second step. The experimental result shows that the algorithm is effective.