Vehicular Mobile Edge Computing is a promising technology to leverage the bottleneck at a base station (BS) at peak hours. However, to deploy Vehicular Mobile Edge Computing requires to deal with the challenges in how to incentive vehicles to resource sharing and how to assign tasks and computation resource to minimize the total network delay. In this paper, we develop a two-stage incentive mechanism and task assignment and resource allocation scheme by combining auction game, matching theory, and convex optimization method. In the first stage, we present the incentive problem between the BS and nearby vehicles, which leverages a reserve auction. Then we study the network delay minimization problem. The problem is decoupled into two subproblems for determining task assignment and computing resource allocation, respectively. Finally, numerical results show the effectiveness and efficiency of our scheme.