The poor operation of charging stations has been clearly listed as one of the major drawbacks for the wide adoption of plug-in electric vehicles (PEVs). Currently, service providers (SPs) of PEV charging are still struggling to make a decent profit, which has caused problems such as poor management of charging stations and degraded experience for PEV users. This article is aimed at exploring the potential of edge intelligence to decide PEV charging pricing strategies under various scenarios, in which the SP's pricing strategies can quickly respond to the dynamic needs of PEV users and load of the grid. First, the key factors and parameters that affect the behaviors and interactions of PEV users, charging SPs, and the grid are introduced. Second, we provide the basic idea of edge intelligence, in particular, how to apply it to vehicular networks. Next, considering the challenges including low sampling rate, large variance, slow convergence, and so on, we discuss the potential of utilizing reinforcement learning algorithms at the network edge to solve the pricing strategy. Moreover, future directions of using edge intelligence for PEV charging pricing strategy are provided.