Quadcopter path planning refers to the quadcopter using an algorithm to design a path that is safe, capable of responding to dynamic changes in the scene, and low-cost under a certain environment, given the quadcopter’s source point and target point, so that the quadcopter can fly from the source point to the target point. The path planning system is an important part of the autonomous control of quadcopters and an important guarantee for quadcopter flight safety. It is also directly related to the quality of quadcopter flight path selection and the smoothness of flight. The traditional path planning algorithm based on gradient descent has problems such as inflexibility and a large amount of calculation. However, learning-based algorithms also lack hard-coded security measures and trajectory constraints, which limits the practicality of learning-based algorithms. This paper proposes a planner, EGO-DQN, that integrates deep reinforcement learning algorithms and EGO-Planner. EGO-Planner is one of the most popular path planners, based on the traditional gradient descent algorithm. EGO-DQN uses reinforcement learning, can adapt to complex environments, has efficient path planning, and has hard-coded safety constraints.