Abstract: Based on urban roads, this paper proposes a layered model predictive control (MPC) to plan car speeds to reduce frequent start and stop of hybrid electric vehicles, improve traffic efficiency, and achieve fuel economy through energy management. The upper-layer MPC uses the intelligent network technology to obtain states like speed and position of the vehicle in front, and predicts the optimal speed of the vehicle to avoid stopping at red lights by combining the traffic environment data. The lower-layer MPC manages the energy of hybrid electric vehicles according to the optimal speed to improve fuel economy. The comparison between the energy management results of the layered MPC and those of dynamic programming at the same speed shows that the proposed method can effectively avoid vehicle idling at red lights and improve fuel economy and traffic efficiency.