According to numerous studies, up to 40% of the total energy is consumed in the building sector. Energy reduction in this sector by means of cost-effective and scalable approaches could have an enormous economic, social and environmental impact. Achievement in substantial energy reduction in buildings may require to rethink the entire processes of design, construction and operation of buildings. As almost half of the energy consumed in this area is used for the heating, ventilation and air conditioning, one possible way to achieve the savings in building sector is to improve the building heating control strategy. A shiny example of the improvement can be the use of model based predictive control (MPC). In this paper, we describe the complete process of the MPC implementation for a real office building in Hasselt, Belgium. Starting with building description and data collection, followed by the discussion about the suitable model structure design and proper identification methods selection, the paper attains to a flexible two-level control concept. Finally, the proposed control scheme leads in average to a 17% energy consumption reduction compared to the conventional control strategy. (C) 2014 Elsevier Ltd. All rights reserved.