Abstract: The flexibility regulation capacity based on energy storage is an important increment for the stable operation of new power systems, and its capacity prediction and allocation are constrained by various factors. To address the problem of difficult to predict the dynamic change of energy storage capacity demand in the context of energy structure transformation, which in turn affects the coordinated planning and construction of source-grid-load-storage, this paper analyzes the influencing factors of power system flexibility regulation capacity, studies the installed capacity of each flexibility regulation resource including thermal power, hydropower, and energy storage as well as the development trend of regulation cost, and carries out segmental linearization of the deep ly peaking cost of thermal power units in order to reduce the complexity of the calculation. Taking the lowest cost of flexible regulation capacity as the optimization objective, the energy and power constraints of each regulation resource are considered comprehensively, and an energy storage capacity prediction model is established. The model is converted to the form of mixed integer linear programming. which is solved by calling the commercial optimization solver CPLEX. The model is based on the regulation capacity of the existing regulation resources in the system and considers that with the increasing proportion of new energy generation, the energy storage capacity is predicted. The analysis shows that the proposed model can make dynamic prediction of the capacity of energy storage under the premise of ensuring the economic operation of the system, and the allocated energy storage can effectively promote the consumption of wind power and photovoltaic.