In order to promote the development of green traffic, this paper develops an intelligent connected vehicle ecological driving system that integrates V2X multi-source holographic traffic perception, vehicle detection and tracking, and dynamic path planning for typical urban road scenes. Firstly, this paper used V2X vehicle-road collaborative technology to collect the traffic flow and the motion state information of the vehicle. The least squares dynamic weighted fusion algorithm is introduced to synchronously process the acquired traffic data. Based on safety distance constraint and kinematic constraint, the ecological driving objective function considering operation efficiency and environmental benefit is constructed. Then, the hybrid A* algorithm and the modified RS curve fusion path planning method are used to solve the optimal driving trajectory of intelligent connected vehicles. Finally, a vehicle-mounted intelligent connected vehicle ecological driving system terminal is developed. The optimal ecological driving track, the optimal speed information and early warning information are sent to the driver in real time. The actual scenario verification results show that the driver can obtain the current traffic state, the warning prompt and the optimal suggested speed through the ecological driving system, which achieves the purpose of improving the driving safety and economy. Through the statistics of the energy consumption of intelligent connected vehicles within a week, we found the energy consumption reduction effect is significant, which shows that the proposed system scheme can provide feasible solutions for the future development of urban green intelligent transportation system.