An approach for monitoring, detecting and localizing anomalies of HVAC system by using the combination of thermodynamics models, the energy balance of a zone in steady state, and data analytics is disclosed. The approach determines, via machine learning, the ideal thermodynamic model for an area serviced by an HVAC system. The approach retrieves reading from various sensors and insert the current sensor reading into the ideal model. In the presence of anomalies, the parameters of the model will deviate from their nominal values and an appropriate action can be taken based on the severity of the detected and localized anomalies.