The Heat Pump with geothermal exchanger is one of the best methods to heat a building. The heat exchanger is an element with probabilities of failure due its size and due it is outside construction. The present study shows a novel intelligent system design to detect faults on this type of heating equipment. The novel approach has been successfully empirically tested under a real dataset obtained during measurements along one year. It is based on classification techniques with the aim to detect failures in real time. Then the model is validated and verified over the building; it allows to obtain good results in all the operating conditions ranges.