Abstract: The decision-making problem for multi-objective optimization is studied in this article. By defining regret value in regret-function and designing the objective function, the priority degree for each object can be quantitatively expressed. The simulation results show that the objective function based on regret-function is of universal use in multi-objective optimization. Compared to the traditional way, it doesnt have to try over and over to find a proper weight. Besides, it can approach the ideal objects value better, and the optimized result can directly reflect designers original intention, which also makes it easier to be improved under different conditions.