Abstract: As the combat missions and environments of unmanned aerial vehicles(UAV) are becoming increasingly complicated, a UAV may not have sufficient resources to complete the assigned tasks. It is necessary that the unmanned aerial vehicles should form a coalition so that they can accomplish the complex tasks more efficiently and improve the success rate as well as effectiveness of tasks in large measure. This paper has provided insight into multi-UAV task allocation algorithm amid military environment. First and foremost, based on the various tasks required by the UAV resources, a resource model is established. Besides, quantitative calculation of various resources required by the UAV alliance was carried out. At the same time, according to the difficulty of the task, this paper proposes a task allocation(TA) algorithm, and also calculates the reward, costs, success rate and effectiveness of each coalition to perform each task. There is no denying that these endeavors will make for selecting the best alliance with the most effectiveness as a way to carry out the task. These efforts have improved the robustness and versatility of the algorithm significantly. In order to verify the effectiveness of the algorithm, simulation experiments are carried out in Matlab. In addition, in a bid to make human-computer interaction more economical and safer, the above mentioned task allocation algorithm, the auction algorithm, leader-follower algorithm and the Hungarian algorithm are compared in success rate and effectiveness. Finally, the experimental data show that the UAV coalition can significantly improve the success rate and effectiveness of task execution.