With the in-depth application of intelligent network technology in the field of transportation, vehicle-road cooperation system has gradually become the core support architecture of intelligent transportation. This paper focuses on the design and performance optimization of cross-layer communication protocols, and proposes a performance-oriented path, including reversible behavior inference model, unified scheduling function framework, event-driven path construction method and semantic abstraction compression mechanism. Then, a multi-scene high-fidelity simulation platform integrating NS-3 and SUMO was constructed to verify the effect of the optimization protocol around traffic patterns such as city intersection and high-speed lane change. The results show that the proposed scheme is highly adaptable to complex environment and can provide reference for the further development of cross-laye communication protocols.