Abstract: Multi-scenario unit commitment model is a typical method to solve the uncertainty problem of power system scheduling. At present, multi-scenario models are built by setting scenario-stage variables. But this modeling approach generates too many constraints and variables, which slows the calculation speed and reduces the practicability in a large-scale power system. In this paper, a new and efficient modeling approach is proposed, and the optimized variables of each node are set through establishing node-stage and node-node incidence matrices. The proposed method reduces the model complexity and accelerates the calculation speed. The case study testifies the feasibility and applicability of the proposed method, and its advantage of calculation speed over traditional method. For large-scale and multi-scenario practical power systems, the node-based multi-scenario unit commitment model has better applicability.