Abstract: With the deepening of China’s electricity market reform process, the first batch of spot electricity pilot areas have completed the work of spot price settlement. However, in the actual settlement process, there is a large number of unbalanced settlement cost, which affects the fairness and efficiency of the market. How to fully estimate the unbalanced settlement cost in the market design stage to reduce the market risk and improve the market efficiency is the direction that market designers need to focus on. The simulation analysis of unbalanced settlement cost is carried out from the perspective of the settlement mechanism in the real-time balancing market and the interaction behavior of market subjects. Firstly, the balancing market elements closely related to the unbalanced settlement cost are extracted. Then, the simulation model of balancing market is built based on agent-based model（ABM） simulation method. The interaction strategy for participating in market transactions is formulated by market subjects using reinforcement learning algorithm. Finally, the impact of the strategic behavior of the balancing service provider, the unbalanced settlement price mechanism, and the tolerance margin of renewable energy on the unbalanced settlement cost is analyzed through the cases.