Multi-UAV path planning is a complex multi-constraint combinatorial optimization problem, which is difficult to obtain good results through traditional algorithms. Recently, the metaheuristic algorithm has become an effective means to solve this kind of optimization problems because of its high efficiency. Firstly, a new classification method is proposed for the key elements of multi-UAV path planning. Then, we classify and discuss the improvement methods of metaheuristic algorithms used in the research results, elaborate the application performance, advantages and disadvantages of these algorithms, and summarize the current research results according to the proposed classification methods and research characteristics. Finally, the future development trend is prospected based on the existing problems in the current multi-UAV path planning research.