In the complicated search environment,multi-UAV cooperative search mission may confront with diverse types of targets.Considering the motion features of different types of target,we used a search map method related to probability density function of the targets,and established a reward function appropriate to the coordinated task in order to improve the efficiency of cooperative search.The control architecture combining centralized with distributed control was used,and multi-UAV cooperative search decision making was implemented through predictive control and improved Hungarian algorithm.The decision-making method could improve the search efficiency while avoiding the possible collision of the UAVs.Finally,contradistinctive simulation results demonstrate the validity of the proposed cooperative search strategy.