The invention provides an adjustable capacity prediction method for an energy storage system of considering multi-source information fusion and planned power output. The method comprises the following steps: (1) obtaining real-time data; (2) calculating a day-ahead predictive power and an ultra short-term predictive power of a new energy power plant at a present moment and a charged state of the energy storage system; (3) calculating a power output value of the energy storage system at the present moment on the basis of an operating state of the new energy power plant and a charged state interval of the energy storage system; (4) distributing the power output value of the energy storage system to various energy storage units, and calculating charged state values of the energy storage units at the end of the present moment; (5) calculating the charged state value of the energy storage system at the end of the next four hours of the present moment; and (6) calculating the adjustable charge-discharge capacity of the energy storage system in the next four hours. The adjustable capacity prediction method is suitable for prediction of the adjustable capacity of the energy storage system during new energy of stored wind, stored light, stored wind and light and the like and stored energy combined generation tracking and generation planning application.