Along with the increasing demand for electricity use in Indonesia, especially north Sumatra, PT. PLN (PERSERO) built a coal-fired steam power plant. With the existence of this steam power plant, it can help the country's electricity supply, especially in Sumatra. As is known that steam power plants are plants that have a high potential for equipment damage, especially the Balance of Plant (BOP) area. Therefore, a team is formed that can overcome and identify equipment damage with certain tools, called the Predictive Maintenance team. Predictive Maintenance (PdM) techniques are designed specifically to help determine the condition of equipment assets used as a reference for predictions for when asset maintenance activities should be carried out. Predictive Maintenance (PdM) is a form of maintenance that directly monitors the condition and performance of equipment during normal operation to reduce damage or failures in the future. The main tools used for Predictive Maintenance are Vibration Meter and Infrared Thermography. And the equipment has standards or limits resulting from trending measurement data and also has international standards as a reference. There was damage to one of the Balance of Plant (BOP) equipment, which is Demineralized (DM) Pump that causes the supply of demineralized water to the unit to be reduced and the unit load cannot be increased to maximum. With predictive maintenance (PdM) analysis, it can be quickly identified that the damage was in the drive end (DE) motor bearing of the demineralized equipment (DM) pump. Replacement of bearing on the Demineralized motor (DM) can maximize the performance of the unit so as to make an electrical energy saving of 1,095,360 kWh. In addition, it also has financial benefits of Rp. 681,781,917,-