To better simulate the irrational decision-making behavior of electric vehicle (EV), this paper proposes an innovative EV charging navigation strategy integrating user regret psychology and deep reinforcement learning considering multiple uncertain factors. Firstly, a mathematical optimization model for EV charging navigation based on user regret psychology is established. It calculates the regret value of strategy by time and charging cost based on historical data. Then, the proposed model is optimized by the Double Deep Q Network (DDQN) reinforcement learning algorithm to minimize the regret value of the strategy. Finally, the proposed method is simulated in an urban traffic network. The simulation results show that the proposed method can effectively solve problems with multiple uncertainties. Compared with the nearest charging strategy based on the Dijkstra algorithm, the proposed strategy can effectively reduce the regret value, thereby lowering the total cost of EV charging, and it is more acceptable to user. Furthermore, the proposed strategy maintains good adaptability and effectiveness in different simulation environments.