At present, the single strategy of aircraft is difficult to penetration success and the decision making mainly depends on the pilot. In this paper, an intelligent aircraft defense decision algorithm based on rainbow algorithm is proposed by combining the two kinds of decision making of aircraft maneuver and decoy projectile. Firstly, motion model and interaction model are established. Then, with line-of-sight and line-of-sight Angle as states, maneuvering and decoy deployment as actions, the reward function considering fuel loss and decoy consumption is designed to establish reinforcement learning environment. The rainbow algorithm in reinforcement learning (RL) is used to train the agent. Considering the requirement of training speed and stability, \(\varepsilon - greedy\) strategy is used to replace the noise network in Rainbow algorithm to speed up convergence and improve stability, and the optimal aircraft intelligent defense decision method is obtained. Finally, the simulation results show that this method can make the aircraft maneuver and deploy decoys autonomously, reduce the dependence on the pilot, and improve the survivability of the aircraft.