Abstract: To improve the effectiveness of operational command decision-making, an intelligent decision-making method for operational command based on reinforcement learning is proposed. On the basis of constructing the combat aided decision-making model based on Q-learning algorithm, the state space, action space and integrated environment in the model structure are designed, and the rules for the changes of the states of the forces and the overall morale of the troops are given, and the principles of the state correction when different actions occur are also defined. Then, the reward function of the decision-making model is designed from the number of annihilated enemies, the loss of forces and the change of overall morale. Finally, the proposed decision-making model is simulated and tested in a classic battle scenario and the classic battle of the Russo-Ukrainian conflict. The experimental results show that the model can input the basic situation of both sides' troops and the comprehensive environment of the battlefield training to obtain the decision Q table, and the decision suggestions based on the Q values of different actions have a strong consistency with the actual combat decision-making, which can effectively assist commanders in making combat decisions.