In the conventional distribution network dispatching mode, the forecasting error and real-time fluctuation are regulated by energy storage and flexible load, and the load values are predicted roughly, which makes the load forecasting often not accurate enough. Moreover, using controllable distributed power supply, flexible load and energy storage device to balance load fluctuation of distribution network and deal with distributed output will result in large flexible load dispatching and reserve cost. A multi-objective optimal dispatching method for active distribution network based on cluster load prediction is proposed. The fuzzy clustering method is used to cluster the load of the active distribution network, and the extreme learning machine is used to predict the load values. Based on the predicted value, the day-ahead dispatching is firstly carried out with the lowest active power dispatching cost. Real-time dispatching was carried out with the objective of minimum controllable distributed output correction, minimum energy storage output correction, and minimum flexible load correction and then modified on the basis of day-ahead dispatching.