In order to improve the economy of power-split hybrid electric vehicle (HEV),a longitudinal dynamics model of the entire HEV vehicle was established,and an energy management strategy (EMS) based on strategy entropy optimization with an improved proximal policy optimization (PPO) algorithm was proposed. The algorithmic framework was simplified by employing an experience pooling mechanism based on traditional PPO algorithm,and only one deep neural network was used for interactive training and updating to reduce the complexity of parameter synchronization in the policy network. In order to effectively explore the environment and learn more efficient strategies,the strategy entropy was added to the loss function to promote the intelligence to strike a balance between exploration and utilization and to avoid premature convergence of strategies to local optimal solutions. The results show that the EMS based on the improved PPO algorithm with single-policy network maintains the state of charge(SOC) of the battery more effectively than the EMS based on the dual-strategy network PPO under both UDDS and NEDC driving cycle. Additionally,the equivalent fuel consumption is reduced by 85% and 14%,respectively,achieving energy-saving effects comparable to the EMS based on the dynamic programming(DP) algorithm.The proposed improved PPO algorithm can effectively enhance the fuel economy of hybrid vehicles and provide a reference for the design and development of EMS for hybrid vehicles.