This paper proposes an implementation of method of heating feed-forward control system for multi-storey buildings. The proposed structure incorporates baseline control loop that controls heating power depending on the key perturbing factor-outdoor air temperature and adjusting control loop implementing feed-forward control based on indoor air temperature inverse dynamics model. The suggested model structure enables real-time assessment of the impact of unmeasurable perturbing factors on indoor air temperature, along with distinguishing between fast and slow processes occurring within the system. To measure current indoor air temperature values, the authors used a distributed field-level sensor network. The estimation of generalized temperature perturbation was performed using predicting ability of exponential smoothing. The paper also contains the actual results obtained by deployment of the suggested heating control system in the university academic building. The results obtained demonstrate the reducing of overall energy consumption by building heating system, at the same time increasing the comfort level of the building.