To solve the problem of multi-UAV cooperative search for dynamic targets, a composite differential evolution algorithm is proposed for multiple UAVs to perform cooperative dynamic target search. First, the environment information graph model for the cooperative dynamic target search and UAV model are established. Then, based on the improved differential bat algorithm and the adaptive differential evolution algorithm, a combinatorial framework relied on population adaptive allocation is designed. By introducing the mutation, crossover and selection mechanisms of the differential evolution algorithm to the bat algorithm, a cooperative search algorithm combined with the differential evolution algorithm is constructed. For the real scenario where the trajectories of dynamic targets are random and unpredictable and have reconnaissance evasion trait, the retrievable digital information graph and the adaptive target search gain function are established to enhance the capabilities of UAVs to capture dynamic targets. Finally, simulation results demonstrate the effectiveness of the proposed UAV cooperative dynamic target search algorithm.