Abstract: Peer-to-peer (P2P) energy trading can encourage prosumers to carry out local energy trading, and guide user-side resources including distributed energy resources (DERs), energy storage, electric vehicles (EVs) and flexible loads to participate in supply-demand balancing through market-based price signals, thereby boosting the local consumption rate of renewable energy. To accurately identify the core research priorities of P2P energy trading, this paper first elaborates its system architecture, operational logic and basic characteristics. Second, it reviews the existing market mechanisms and their applicable scenarios from the perspectives of market organization models, pricing, market clearing and settlement mechanisms. Third, it summarizes state-of-the-art research on strategy optimization approaches for P2P energy trading, covering mathematical optimization, game theory, distributed optimization, reinforcement learning and other methodologies. Finally, the paper outlines future development directions of P2P energy trading targeting critical challenges such as physical constraints of distribution networks, large-scale agent coordination, privacy protection, trusted trading and engineering demonstrations. The work is expected to provide references for market-oriented trading of user-side distributed resources and the operation of new distribution systems.