Desiccant wheels are commonly used to dehumidify air in air-conditioning system to reduce the energy consumption. The objective of this study was to design, test and analyse the performance of a novel tube-shell, internally water-cooled desiccant wheel. Cooling water was used in the supply section of the new wheel to shift the dehumidification process from nearly adiabatic to nearly isothermal. We conducted a series of experiments to investigate the wheel's performance and used the experimental results to validate a mathematical model. The model was then applied to investigate the influence of modifications to the base design, including changing the number of desiccant layers inside the heat-exchange tubes, and the size and number of tubes. We also analysed the influence of water temperature and flow rate on both dehumidification and temperature rise across the wheel. Our results show that isothermal dehumidification performance can be achieved, but only with no more than two desiccant layers inside the heat-exchange tubes. More layers reduced heat transfer between the air in the innermost layers and the cooling water. Lower cooling water inlet temperatures led to lower air outlet humidity and temperature. A cooling water temperature of approximately 24 degrees C was required to achieve isothermal dehumidification for process air with inlet temperature of 30 degrees C and absolute humidity of 16.3 g/kg and regeneration air with the inlet temperature of 50 degrees C and absolute humidity of 16.3 g/kg. This corresponds to a 48% improvement in dehumidification performance (the maximum absolute humidity change between inlet and outlet process air) compared with a conventional adiabatic desiccant wheel while using super-adsorbent polymer as the desiccant material.