Aiming at the problem of multi-UAV cooperative searching for multiple moving targets in the area, considering the detection and false alarm probability of sensors, the flight and collision avoidance constraints of multi-UAV, and the random motion of targets, a multi-UAV three-dimensional cooperation search method based on an information map is proposed. Taking the balance of short-term benefits, long-term benefits and coordination benefits of UAV search as the core, considering the characteristics of the three-dimensional motion of a UAV, a mathematical programming model for multi-UAV cooperative search is constructed, and a search information map is designed with four factors of existence probability, environmental uncertainty, revisiting pheromone and searching gain. The model is solved based on the rolling planning architecture, integrating the newly proposed pruning method. In a typical cooperative search scenario, the effectiveness of the proposed method is verified by numerical simulation experiments. The simulation results show that the proposed method can make the three-dimensional flight decision of each UAV in seconds, revisiting pheromone and searching gain factors can guide the UAV to capture more targets. The comparative simulation results show that the proposed method can capture more targets while having fewer misjudgments, which effectively improves the task efficiency of multi-UAV cooperative search.