With the popularity and application of Internet-of-Vehicles (IoV) , vehicles need to accomplish a large number of real-time computational tasks. To enhance the wireless connectivity and computational capability of the vehicles, the introduction of unmanned aerial vehicle (UAV) with the advantages of high mobility and on-demand deployment for assistance is an effective method. Therefore, this paper studies an air-ground collaborative mobile edge computing system for the IoV, which consists of mobile edge computing servers deployed at a UAV and ground base stations, respectively, to collaborate in providing communication and computation services for vehicles. In order to minimize the maximum average communication and computational delay of the vehicles, this paper studied a problem of jointly optimizing communication bandwidth allocation between the UAV and ground base stations, computational task offload ratio allocation, UAV trajectory and computational resource allocation. To solve this non-convex optimization problem, this paper proposed an efficient alternating optimization algorithm based on block coordinate descent and successive convex optimization methods, which is decomposed into four sub-problems: bandwidth allocation, computational task offload ratio allocation, UAV trajectory optimization and computational resource allocation, and introduced slack variables and utilized a first-order Taylor expansion to solve the sub-problems alternately and iteratively. Simulation results showed that the proposed algorithm can effectively reduce the average communication and computational delay of the vehicle compared to the multiple benchmark schemes. This demonstrates the importance of air-ground collaboration between the UAV and ground base stations to enhance the communication and computation capabilities of the IoV.