The continuous expansion of renewable energy sources and their ongoing integration into existing power networks, alongside the emergence of new types of loads, has led to signifcant new challenges emerging for power grid source–load coordination scheduling in recent years. To fully harness the demand response (DR) potential of the load side of such systems, this paper proposes an optimal scheduling method based on DR and the Stackelberg game. A Stackelberg game model is established to formulate day-ahead source–load coordinated scheduling problem, taking account of the stochastic dynamic characteristics of the load response process and user electricity consumption satisfaction with electricity consumption. The model is essentially a strongly nonlinear optimization problem, however; thus, to avoid several of the limitations seen in traditional optimization methods, a data-model fusion-driven method is proposed, with electricity prices formulated by the data-driven method used as input for a model-driven method whose output is then used in turn to guide a data-driven formulation of electricity prices. Case studies on IEEE test systems are used to verify the feasibility and accuracy of the proposed method.