Abstract: In order to obtain higher earnings, power suppliers or agents hope to enact flexible adjustable electricity price by use of peak-valley price and the prices under other load situations to promote the flexible interaction among intelligent residential areas and distribution network. The premise of such a scheme is not increasing the user’s total cost of electricity consumption and their demand on the satisfaction degree should be met, however, the change and discrepancy among different consumers’ satisfaction should even more be considered. For this reason, a bi-level model of smart community agent pricing and smart electricity service plan, in which the change and difference of consumers’ satisfaction degree were considered, was proposed. In the upper level of this model the maximized profit of the agent was taken as the objective and the user electricity price was enacted and transmitted to the lower level. In the lower level of this model the lowest users electricity charge was taken as the objective and considering the constrains of change of user’s satisfaction and difference the smart power utilization plan was generated and sent back to the upper level, and the proposed bi-level model was solved by the iteration of upper and lower level mixed integer programming. Results of computing example show that after taking the change and difference of satisfaction into consideration, the earning of the agent is increased and the user’s electricity charge is decreased, and the difference between peak and valley in the load curve is decreased as well. The pricing of smart community agent can adjust the flexibility of power supply market, improve the characteristics of power load peak-valley difference curve, and provide user more personalized smart electricity service after considering the change and difference of user satisfaction.