Configuring a certain capacity of energy storage for the power system can effectively improve the reliability of the power supply and the level of wind power consumption. This paper takes pumped storage investment cost and wind power consumption demand as the optimization goal, realizes the coordinated operation of pumped storage units and thermal power units, and considers the uncertainty of wind power and load, the multi-timescale characteristics of different types of units and load demand response, and establishes a multi-timescale pumped storage capacity optimization model based on stochastic programming. The model consists of inner and outer layers. The outer layer is investment decision-making, which decides the allocation of pumped storage capacity. The inner layer is for optimization decision-making, and the system optimization operation is carried out in three stages: day-ahead short-term optimization, intraday ultra short-term optimization, and real-time optimization. An example of a typical scenario is established, and the planning results verify the effectiveness of the proposed model and method. The thermal-wind-hydro power system's total cost decreases first and then increases with the increase in the installed capacity of the pumped storage, and the curtailment rate gradually decreases. The total cost increases faster when the pumped-storage installed capacity is larger than optimal. For a pumped-storage power station of the same capacity, variable-speed pumped storage is better than fixed speed pumped storage in reducing the wind curtailment rate. The main reason is that its output is continuously adjustable under the condition of the variable-speed water pump, which is especially suitable for energy storage at night when wind power is high.