Accurately describing the error characteristics of wind power output prediction is helpful for the rational allocation of system reserve capacity and the optimization of day-ahead scheduling plans. This paper proposes a day-ahead optimization scheduling method for power systems considering wind power ramping reserve requirements. First, based on the wind power ramping segment, the ramping characteristics are extracted, and a two-dimensional interval of ramping amplitude-predicted power is established. The adaptive kernel density estimation method is used to fit the probability distribution of wind power prediction errors. Then, based on the distribution of wind power prediction errors, the system reserve requirements are determined. Looking for minimal comprehensive operating costs of reserve costs and risk costs, a continuous-time day-ahead optimization scheduling model is established. Next, the Bernstein polynomial interpolation solution space transform is adopted to complete model conversion, thereby optimizing reserve capacity, unit combination, and output plans. Finally, a case study verifies that the established wind power prediction error distribution model can accurately describe the stochastic characteristics of wind power. The proposed day-ahead scheduling method can effectively allocate system reserve capacity, ensuring operational safety and economic efficiency.