Abstract The demanding computing capacity of emerging vehicular applications has emerged as a challenge in Internet of vehicles (IoVs). Multi-access edge computing (MEC) can significantly enhance computing capability and prolong battery life of vehicles through offloading computation-intensive tasks for edge computing. Considering the impact of vehicles’ mobility on communication quality, this paper provides an energy-efficient computation offloading scheme for vehicular edge computing networks (VECN). An energy-efficiency cost (EEC) minimization problem is formulated to make a tradeoff between latency and energy consumption, for completing computational tasks in an effective manner. Since that multiple variables and time-varying channel conditions make the formulated problem difficult to solve, we transform the original non-convex problem into a two-level optimization problem and develop an iterative distributed algorithm to obtain an optimal solution. Numerical results verify the convergence and superiority of the proposed algorithm.