In the centralized scheduling of multi-residents, the complexity of scheduling time will be greatly increased as the number of resident increases. In order to reduce the time complexity caused by centralized scheduling, a probability model based on the Time-of-use electricity tariff diff erence is proposed and applied to distributed estimation of the algorithm. According to the impact factor mechanism of the probability model of Time-of-use electricity tariff diff erence, not only the time complexity of centralized scheduling is reduced, but also the optimization of the algorithm will not fall into a local optimal situation. In the centralized scheduling model of building residents, the controllable load of residents and new energy are centralized. The consumption rate of new energy was improved by changing the new energy power supply mechanism. Under the conditions of ensuring the comfort of household electricity consumption, three objective functions of the model include: (a) to reduce the total daily electricity consumption, (b) to fl atten the peak-to-valley diff erence of daily electricity, (c) to decrease the discarded rate of new energy. The simulation of the calculation example verifi es the feasibility and eff ectiveness of the proposed method.