Air-Gap Membrane Distillation (AGMD) promises to reduce heat loss in membrane distillation. Most AGMD models are one-dimensional and do not consider the downstream variations. In addition, a linear function of vapour pressure is used, which either relies on experimentally determined parameters or a simplified mass transfer resistance to model the water permeate flux. This study introduces a new, improved model that simultaneously considers both heat and mass transfer in the AGMD process by coupling the continuity, momentum, and energy equations. A novel precise logarithmic function of vapour pressure was derived to model the water permeate flux, independent of experimentally determined parameters. By varying the inlet temperature, Reynolds number, inlet concentration, and air-gap thickness, the performance of AGMD was evaluated. The results revealed that our model improved the water flux prediction from more than 10% to less than 4% deviation from experimental results. Among the operating conditions, only increasing the Reynolds number improved all the system performance metrics, including higher water flux and lower temperature and concentration polarisation effects. Results were compared with Direct Contact Membrane Distillation (DCMD) outcomes and showed that unlike AGMD, DCMD suffers from a substantial decrease in water flux along the module. For DCMD, the exit water flux value decreased by 50% in comparison with the inlet value, while the water flux decreased by only 2% for AGMD, using a 1 mm air gap thickness.