In the past few decades, as a prominent representative of the third generation of intelligent robots, drones have taken the lead in developing from single task to multi machine collaborative work. Research has shown that using multiple drones to search for target areas can expand the search field, improve the efficiency of drones in completing tasks, reduce energy consumption during flight, and reduce environmental instability. Thus, high-precision positioning can be further carried out to improve the target hit rate. However, clustered drones also face some new challenges, such as the robustness of multi-agent topology, real-time performance of online perception mechanisms, reliability of information exchange, and loose coupling of collaborative systems. The method proposed in this paper improves the problem that the traditional Ant colony optimization algorithms is easy to fall into the local optimum and the Rate of convergence is slow. The obstacle avoidance factor is added to the calculation formula of the state transition probability, and an improved Pheromone volatilization coefficient based on the Gaussian distribution is given, so that the Pheromone volatilization factor changes from a fixed value to an adaptive value that changes with time, which makes the path obtained by the algorithm better and greatly speeds up the Rate of convergence of the algorithm. In this system, all UAVs can share their original data through direct communication between them. Even if one UAV fails, the whole task will not be affected.