Peer-to-peer (P2P) energy sharing involves novel technologies and business models at the demand-side of power systems, which is able to manage the increasing connection of distributed energy resources (DERs). In P2P energy sharing, prosumers directly trade energy with each other to achieve a win-win outcome. From the perspectives of power systems, P2P energy sharing has the potential to facilitate local energy balance and selfsufficiency. A systematic index system was developed to evaluate the performance of various P2P energy sharing mechanisms based on a multiagent-based simulation framework. The simulation framework is composed of three types of agents and three corresponding models. Two techniques, i.e. step length control and learning process involvement, and a last-defence mechanism were proposed to facilitate the convergence of simulation and deal with the divergence. The evaluation indexes include three economic indexes, i.e. value tapping, participation willing and equality, and three technical indexes, i.e. energy balance, power flatness and self-sufficiency. They are normalised and further synthesized to reflect the overall performance. The proposed methods were applied to simulate and evaluate three existing P2P energy sharing mechanisms, i.e. the supply and demand ratio (SDR), mid-market rate (MMR) and bill sharing (BS), for residential customers in current and future scenarios of Great Britain. Simulation results showed that both of the step length control and learning process involvement techniques improve the performance of P2P energy sharing mechanisms with moderate ramping/learning rates. The results also showed that P2P energy sharing has the potential to bring both economic and technical benefits for Great Britain. In terms of the overall performance, the SDR mechanism outperforms all the other mechanisms, and the MMR mechanism has good performance when with moderate PV penetration levels. The BS mechanism performs at the similar level as the conventional paradigm. The conclusion on the mechanism performance is not sensitive to season factors, day types and retail price schemes.