The invention discloses a decision-making method for large-scale unmanned aerial vehicle (UAV) cluster dynamic confrontation. The method comprises the steps that UAVs conduct situation assessment according to environment information of a battlefield and make multi-UAV cooperation and multi-target distribution decisions with maximizing attack benefits as the target; a cluster motion model based onthe society is established, the UAVs select corresponding behavior rules according to the target distribution result to update the speeds and locations; the UAVs make attack decisions and judge whether distributed targets meet the attack conditions or not, if yes, the UAVs attack the corresponding targets and update the number of weapons and the survival probabilities of the enemy targets, and ifnot, the UAVs judge whether the military operation is ended or not. According to the decision-making method, the UAVs conduct target distribution and negotiation through communication in real time; the cluster motion model is established to achieve updating the speeds and locations of the UAVS; and accordingly, dynamic confrontation between UAV clusters is further realized.