In urban water supply systems, pump stations are the hubs for making the complete systems operate regularly as well as the main energy-consuming units. In order to address the current problems of water supply systems, such as high energy consumption and low efficiency of the pump station operation, and poor response and adaptability to disturbance events, a digital twin (DT)-based full-process dynamic pump station scheduling method for energy-saving optimization in water treatment plants was proposed in this study. To be specific, the DT technology was introduced to predict the availability status of the pump unit in advance, trigger the rescheduling process in time, and achieve energy conservation and consumption reduction, so as to provide technical and methodological support for unattended pump stations. The results of experiments revealed that an average energy-saving rate of 9.78% could be achieved by using the proposed method on the premise of ensuring the full-process dynamic water balance. In addition, the method could maintain high efficiency during the operation of the pumps, and guarantee the safety and stability of the pump stations.