Aiming at the aircraft control system design problem independent of accurate model, the intelligent control algorithm based on deep reinforcement learning is mainly explored. First, the flow chart of Twin Delayed Deep Deterministic Policy Gradient(TD3) algorithm is given, then the aircraft control problem is modeled as Markov decision process, the action-state space and reward are designed, and the deep neural network structure and its training process of the agent are described. Finally, digital simulation analysis is carried out. The trained Reinforcement Learning(RL) controller is compared with the traditional Linear Quadratic Regulator(LQR) controller. Simulation results show that the RL controller designed in this paper can achieve the expected control objectives, has certain robustness and better control performance than LQR, thus proving the effectiveness of the proposed control strategy, laying a foundation for further engineering applications.