In order to improve the lack of reality and flexibility of the aim path-planning in traditional approximation method, a new method of formation change based on genetic algorithm was presented. By reducing the entities' sequences difference between initial formation and final formation, the chromosome coding and fitness function were designed. Furthermore, the algorithm was applied to a CGF simulation system. The experimental results showed that the proposed algorithm was indeed an effective way to solve the problem of aim path-planning in formation change. Compared with existing algorithms, this algorithm had more reasonable result and the possibility of collision was significantly reduced.