With the continuous increase in residential electricity consumption, the potential for demand-side flexible loads to participate in demand response (DR) has become significant. To address the scheduling challenges of flexible loads in residential communities participating in DR under residents’ bounded rationality, an energy usage model of household appliances, a stochastic user behavior model, and a comfort value model are constructed. Combined with relevant residents’ travel and weather forecast data, time-of-use pricing, and DR policy information from the grid, a day-ahead stochastic optimal dispatch model is developed based on prospect theory (PT), generating day-ahead scheduling plans. This paper provides reliable data for optimizing dispatch in residential communities and realistically simulates residents’ decision-making psychology in DR participation. Finally, the model is validated in a real residential community in Gansu. The results show that the bounded rationality of decision-makers significantly affects the optimal scheduling of residential communities, and that the PT-based day-ahead stochastic load optimization model can maximize overall benefits while ensuring user comfort.