Ground heat accumulation caused by imbalanced heating and cooling loads in a building can cause the heat-source temperature to increase as the operating age of a geothermal heat pump (GHP) system increases. An alternative system to improve upon this situation is the hybrid GHP system. This study reviews existing research on GHP systems and hybrid GHP systems, variable water flow (VWF) control, and coefficient of performance (COP) prediction. Generally, constant flow control is applied to the circulating pump to provide a flow rate according to the maximum load. The need for VWF control was identified because the hybrid GHP system is used mainly as a heating and cooling heat source system for partial loads rather than the entire building load. Previous studies on predicting the COPs of GHP systems developed prediction models by selecting input values based on mathematical models, collecting data through multiple measurement points, and utilizing data from production environments. The model can be limited by the field environment, and it is necessary to predict the COP using machine learning based on existing field monitoring data.