An improved algorithm is presented based on the GVF Snake in order to achieve the multi-objective edge extraction. Firstly, the GVF Snake is used for carrying out the edge extraction of the whole object region, and the results will contain false edges because of the stagnation point region. Secondly, this false edge is analyzed and a new initial contour is established automatically. Then, the GVF Snake is used again to carry out new edges extraction and to determine the real edge and false edge. In the end, those real edges are connected, which were extracted in above two steps, and remove the false edges to get the complete multi-objective edge. Experimental results show that the proposed algorithm can achieve the multi-objective edge extraction.