Variable refrigerant flow (VRF) system is a popular building air conditioning system which could provide cooling or heating to individual rooms independently. The system is called “variable refrigerant flow” system due to its capability of regulating the refrigerant flow via the precise control of variable speed compressors and electronic expansion valves in each indoor unit. In this dissertation, an advanced VRF system which could provide space cooling, heating and water heating is experimentally evaluated in cooling and heating season for both heat recovery operation and water heating operation. The VRF system is simulated in EnergyPlus and validated with experimental data. Based on the deviation analysis and literature review, it is found that the existing VRF model could not fully reflect the operation characteristic of VRF systems, leading to a high uncertainty in cooling/heating energy and energy consumption. A new VRF model is thereafter proposed, validated in this research and resulted in a model uncertainty less than 5%. Based on the new model, the seasonal performance of an energy saving control strategy and the concept of chilled water storage are investigated. Meanwhile, to solve the mismatch between the building’s thermal load and cooling/heating capability of the VRF system, a new VRF system with phase change material (PCM) based thermal energy storage (TES) is proposed. The new VRF system utilizes single TES device to support both cooling and heating season operation. The performance of new VRF system with PCM based TES is investigated and compared to that of the baseline VRF system. It is found that the new VRF system with PCM based TES could achieve both energy efficiency and demand response goals in cooling and heating season. Based on the comparison, the effect of operation strategies and grid incentive program are discussed. Finally, the economic analysis of the new VRF system with PCM based TES based on annual performance is carried out.