Aiming at throughput and fairness issues for multi-unmanned aerial vehicle (multi-UAV) as aerial base station auxiliary communication,a power allocation algorithm based on the multi-agent deep deterministic policy gradient (MADDPG) algorithm is proposed.The algorithm improves system throughput and fairness by jointly optimizing the power allocation and user access of UAV base stations.This paper firstly constructs a 3D scene in which the UAV base station establishes communication services for the ground and then constructs a problem model of maximizing throughput and fairness by combining jonit power,user association,and UAV location constraints.Considering the complexity of the problem,this paper models the constructed optimization problem as a Markov decision process (MDP),and solves the problem by MADDPG.The simulation results show that compared with other algorithms,the UAV base station power allocation algorithm based on MADDPG proposed in this paper can effectively improve the throughput of the system and the fairness of users,to improve the service quality of communication.