Cloud computing technology, as the foundational infrastructure for rail transit systems, offers substantial computational power to support their operations. This work addresses the current challenges of urban rail transit operation (RTO) and proposes a city-oriented RTO scheduling coordination framework utilizing cloud computing technology. The collaboration between RTO and scheduling is studied using cloud computing, and a collaborative optimization method of RTO and schedule based on a cloud-based neural network is proposed. Experimental results show that the time consumption of using for rail transit vehicle scheduling is significantly lower than that of a traditional algorithm, which clearly demonstrates the advantages of using this algorithm in congested path scenarios. The research results of this paper can provide a reference for improving the rail transit system operation.