This paper presents a case study to develop optimal control strategies of a heat recovery system (HRS) of an existing large factory building. The target building operates multiple chillers even in winter to handle heat generated from manufacturing facilities, while simultaneously providing heated air for occupants. The HRS utilizes cooling water (35–40 ℃) from the chillers’ condensers to increase the water temperature (30–35 ℃) entering the heating coils. In order to develop optimal control strategies of the HRS, a physics-based simulation model was introduced to address the interwoven dynamic relationships between chillers, cooling towers, steam boilers, heat exchangers, and pumps. In other words, the simulation model was constructed in a federated fashion where all relevant inputs and outputs are interconnected to one another. Subsequently, two control strategies of the HRS were derived, optimal vs. simplified rule-based controls. It was found that the optimal control could save energy by 90.3% in the month of January (existing control: 1,378.0 MWh, optimal control: 134.6 MWh, simplified control: 138.4 MWh), while the difference between optimal vs. simplified controls exhibits marginal.