Abstract: When the objective function structure of constrained optimization problem is relatively complex and the constraint conditions are relatively harsh, the convergence performance of Differential Evolution (DE) is worse. In order to take advantage of swarm intelligence search algorithms, a Multi-Strategy fusion Differential Evolution (MSDE) based on hierarchical division, state transition and constraint processing is proposed in this paper. Firstly, the parent population is classified into different grades according to the values of the objective function and the constraint violation degree, and the first grade is divided into three levels according to the hierarchical characteristics. Then, the evolutionary operation is designed by using the characteristics of different grades and hierarchies, this process can improve the exploration capabilities of the differential evolution. Next, the state transfer of the population is carried out, so that the transferred individuals have a better distribution in the decision space, and the constraint processing technology is designed according to the distribution characteristics of the infeasible solutions, which can improve the probability of individual convergence to the feasible domain, and make the infeasible solution converge to the feasible domain as quickly as possible. Finally, the performance of MSDE is stand out by simulation with four state-of-the-art algorithms, and the experimental results show that the proposed strategy improves the performance of the algorithm.