Small/Medium commercial buildings, such as retail stores, restaurants and factories, typically utilize multiple roof top units (RTUs) to serve space heating/cooling. For these buildings it is difficult to apply model-based predictive (MPC) algorithms due to lack of sensors and a building management system (BMS). Furthermore, buildings with open spaces served by multiple RTUs can have significant spatial comfort variations and several RTUs that can be coordinated. The conventional approach for these buildings relies on local feedback control where each RTU is controlled by its own thermostat. This often leads to poor coordination between the RTUs and issues on a frequent compressor cycling with opportunities for demand reduction and efficiency improvements. Due to the practical difficulties, there are very few advanced control algorithms that have been developed for these types of buildings. Our goal is to develop a practical control algorithm for multiple RTU coordination targeting reduced energy consumption and cycling with low sensor requirements. We also introduce a modeling approach, a coupled reduced-order CFD model, which was developed to capture the spatial comfort variations in open spaces. The overall control algorithm is tested and analyzed using the simulation platform. The modeling tool is used to show the effects of changing thermostat locations on thermal comfort and building energy consumption for the RTU coordinator. The paper demonstrates that the coordination algorithm captures the integrated building system characteristics associated with RTU performance and supply/return duct distribution with limited sensor information and can achieve significant energy savings. Compressor short cycling is also reduced as a result of a simple heuristic strategy.