This study presents a model-based control strategy for a novel dedicated outdoor air-chilled ceiling (DOAS-CC) system with the aim of optimizing the overall system performance. The DOAS-CC system incorporates liquid desiccant dehumidification and membrane-based total heat recovery technologies. Simplified but reliable models of major components in the DOAS-CC system are firstly developed to predict the system performance. A cost function is then constructed to minimize total energy consumption while properly maintaining thermal comfort reflected by indoor air temperature and relative humidity. Genetic algorithm is used to search for optimal set-points of the supply air temperature and humidity ratio of the dedicated outdoor air subsystem as well as the supply water temperature. The performance of this strategy is tested and evaluated with different control settings in a simulated multi-zone space served by the DOAS-CC system under various weather conditions. The results show that optimized control variables produced by the optimal strategy can improve the system energy performance and maintain indoor thermal comfort. (C) 2011 Elsevier Ltd. All rights reserved.