The Internet of Vehicles (IoVs), composed of smart cars, roadside nodes, base stations, and the like, has emerged as one of the most prevailing research fields. The IoVs constitutes a delay-sensitive context that requires low-delay services; yet, the current quality of routing services is unassured. The cloud-edge collaboration solution possesses powerful computing and storage capabilities along with relatively short transmission latency, and can fulfill the diverse requirements of users to the maximum extent. This paper conducts research on the routing problem within the IoVs assisted by cloud-edge collaboration to further reduce the service delay of the IoVs. This paper presents a vehicular network routing scheme based on the DQN algorithm. The service nodes make routing decisions and content caching decisions for each request by the user request information and the current network resource status within the system. The proposed strategy attains the goal of reducing network latency and improving service quality by jointly optimizing the allocation of computing, caching, and communication resources. Simulation results indicate that the proposed strategy performs better than the existing cloud-edge collaboration solutions and converges rapidly in network environments with diverse parameters.