Abstract: Based on Pareto dominance, this paper proposes a two-stage multi-objective optimization algorithm for two-dimensional and three-dimensional multi-objective problems. In the global search stage, the population is sorted according to the Pareto dominance relation, and the corresponding selection strategy is carried out according to the ranking level of the critical layer subset. In the local adjustment stage, the individuals in the population are fine-tuned. The new obtained individuals are compared with the nearest individuals in terms of dominance, distribution and convergence, and then, the poor individuals are replaced. The effects of the two stages on the performance of the algorithm are analyzed, and the locally adjusted population is compared, whose results show that the local adjustment strategy can effectively enhance the algorithm performance. By solving the standard test function and comparing with other classical multi-objective algorithms, it is verified that the proposed algorithm can attain better convergence and distribution.