Abstract: According to the existing microgrid "source-charge" bilateral uncertainty problem under the background of Active Distribution Network (ADN), the microgrid load optimization scheduling model is optimized. Firstly, according to the user uncertainty factors under the comprehensive demand response, the user load prediction model and the bilateral interaction model of supply and demand in the microgrid are improved; Secondly, based on the incentive demand response to user uncertainty, the incentive impact parameters are introduced to improve the microgrid user satisfaction and power supply side cost model. Finally, the model is solved using an improved chaotic particle group algorithm. The example results show that the model reduces the power supply side cost of microgrid, improves the user satisfaction, and maximizes the linkage benefit of supply and demand of microgrid.