To address the impact of the uncertainty of renewable energy output on scheduling plans, a robust scheduling method considering stochastic scenarios of renewable energy is proposed. Firstly, for the generation of stochastic scenarios of renewable energy output, the interval and temporal characteristics of errors are taken into account. Kernel density estimation and Markov chain modeling are employed,followed by an improved K-means algorithm for scenario reduction, to generate day-ahead stochastic scenarios of renewable energy output for computation. Secondly, to tackle the issue of intraday deviations of renewable energy output from predicted values, a robust scheduling model incorporating stochastic scenario constraints is constructed, which considers both stochastic scenario constraints and scenario transition constraints. Furthermore, due to the large number of stochastic scenarios leading to an oversized robust scheduling model, a solution method based on stochastic scenario feasibility verification is proposed. Finally, the economic efficiency, safety, and effectiveness of the proposed robust scheduling model and solution method are validated through case studies based on a provincial and regional power grid framework. The results demonstrate that the proposed robust scheduling model can effectively solve the unit commitment problem in power markets with high penetration of renewable energy, and the running time and solution efficiency of the proposed model’s solution method are acceptable in practical market applications.