The image segmentation effect of balloon force Snake model largely depends on the initial parameters' selection. A new method based on Genetic Algorithm (GA), which is efficient, parallel and global searching, was proposed to solve the selection of optimal parameters. In this paper, the parallel genetic computation was used to calculate optimal parameter, the energy function of Snake was used as an object function, and the image similarity function was used as the criteria to stop genetic iterating. The results of real medical images prove that the proposed method can avoid the trivial of selecting parameters artificially through a large number of experiments, also solve the problem of not ideal result caused by unsuitable parameters' values, and it can get excellent segmentation effect.