Due to the advantages of low cost, intelligence, and strong adaptability, Unmanned Aerial Vehicle (UAV) technology has developed rapidly. The development of heterogeneous UAV strategies with high autonomous capabilities has become the main development direction of UAVs. Heterogeneous UAVs have performance differences and asymmetric combat problems where the performance of a single unit is inferior to that of the opponent. It is necessary to cooperate appropriately to improve the performance of the UAV swarm. This paper proposes a conflict search-based strategy for heterogeneous UAV pursuit, which combines a double-layer selection matrix and a time window to plan occupancy based on known enemy information, and uses extended search and conflict resolution to screen feasible solutions in the double-layer matrix. Experimental results show that compared with particle swarm optimization and adaptive target prediction, the success rate of the algorithm is improved by 40.3% and 16%, respectively.