Based on the reinforcement learning algorithm， an event-triggered distributed adaptive optimal control method is proposed for swarm unmanned aerial vehicles （UAVs） with input delays and external disturbances. First， the neural network-based reinforcement learning algorithm is introduced for the optimal control. Second， a dynamic event-triggerred strategy with system control performance is designed. This strategy aims to reduce the communication resource waste and the negative effects on consensus control performance， and to eliminate the Zeno behavior. Third， a coordinate transformation with integral terms is introduced to compensate the input delays. Under the influence of the input delays and external disturbances， the proposed optimal distributed cooperative neural network control strategy based on disturbance observers can ensure that all signals of each UAV are bounded， and the output of each UAV system can achieve consensus. Finally， simulation results verify the effectiveness of the proposed control method.