After a new round open of electricity sales market, a large number of new sell electricity company involved in the sale of electricity market. Which leads to market instability. Because hair, sale members on both sides of the market in the electric power business behaviour real-time change. So only through the analysis of theory and can’t get the real power market operation state. Based on the analysis of electric power market supply and demand ratio, power suppliers and electric power company in different than the market price of supply and demand, set up to simulate long-term market members under supply and demand than bidding decision-making model, the application of RE - Learning, reinforcement Learning algorithm in 10 power suppliers and 30 sell electricity companies bidding behaviour simulation market under the participation of members. The experimental results show that the use of bidding decision model is a good way to express market members under different supply and demand than market clearing. Which verifies the effectiveness of the proposed method.