It is difficult to consume electricity when renewable energy output is large. And it is also difficult to secure the power supply when renewable energy output is small, which is the difficult problem in China. Increasing the flexibility and adjustment ability of the power grid is the main means to solve this problem. In the paper, an optimization method for peak-shaving capacity and adjustment ability of provincial power systems considering multiple uncertain factors is proposed. It is characterized that flexible resource in all types of the electric power system dispatching is carried out. The mutual supporting ability can be improved and renewable energy can be consumed well by provincial tie-lines. Firstly, a flexible resource scheduling model considering power supply, network and energy storage is established. The flexibility of the power system is improved by full use of thermal power units, provincial tie-lines, and energy storage participation in the peak-shaving of the power system. Then, considering multiple uncertainties in the system, an optimization model of provincial power system peak-shaving capacity demand is established to obtain the best peak-shaving capacity with the lowest expected cost. In addition, the day-ahead operation scenario of provincial power system is generated based on Latin hypercube sampling, and the scheduling results of various flexible resources under different scenarios are solved based on day-ahead economic scheduling. Finally, a power system is taken to prove the effectiveness of the proposed method.