This article studies the purchaser’s bidding strategy in bilateral transaction power market. In response to the one-on-one negotiation problem between power purchaser and generator, fuzzy probability is applied to describe the asymmetry of market information, and the Nash equilibrium of both negotiating parties under fuzzy probability and the optimal purchaser’s bidding strategy at this time are studied. Furthermore, utilizing Agent intelligent learning to conduct multiple rounds of price negotiations, and applying fuzzy probability adjustment learning to negotiation information to optimize pricing strategies. An example analysis has demonstrated that this method is effective for energy purchaser to participate in contract negotiations.