Abstract: This paper presents a new method to design regional coverage satellite constellation, the non-dominated sorting genetic algorithm Ⅱ （NSGA-Ⅱ） based on Pareto optimal is improved and applied it to the optimization of regional coverage satellite constellation. The best solution,depending on the importance of different objects, is selected by a kind of multi attributes decision making method. Simulations on remote sensing satellite constellation are presented. The results of the simulation realized by STK and Matlab show that the algorithm can get a group of Pareto solutions. The algorithm presented in this paper can avoid selecting weights of multiple objects. On the other hand, compared with the simple genetic algorithm, our algorithm is more active. Thus the question of optimization of satellite constellation with multiple objectives can be solved.