With the increasingly prominent problem of fossil energy, wind power as a representative of clean energy has developed rapidly, but the uncertainty of wind power output poses a huge challenge to the safe and stable operation of the power grid. Firstly, this paper analyzes the uncertainty characteristics of wind power, and it is found that the forecast error of wind power has two characteristics of time sequence and interval. On this basis, according to the characteristics of forecast errors, this paper improves the Markov chain, and proposes a day-ahead scenario generation model of wind power based on the improved Markov chain, which can effectively describe the sequence and interval characteristics of forecast errors, build an effective and accurate day-ahead scenario set. The results show that the day-ahead wind power scenario generation method proposed in this paper is more reasonable than the wind power scenario generation by traditional methods, and it can incorporate more uncertain features. Finally, this paper takes the minimum probability distance as the objective to reduce the initial scenario set to obtain the typical scenario set, so as to improve the efficiency of stochastic optimization.