We developed a computational fluid dynamics (CFD) model that can describe the transport phenomena and performance in a vacuum membrane distillation (VMD) module for ammonia−water separation. This model is based on a multicomponent approach that considers the flow, heat transfer, and mass transfer for the liquid phase on the feed side and gaseous phases on the permeate side. The permeation fluxes across the membrane were modeled theoretically based on Knudsen diffusion and Poiseuille flow. A calibration parameter was introduced to compensate for the error in the permeation flux caused by the uncertainty in the membrane properties necessary for the calculation of the permeation flux. To validate the model, we conducted a laboratory-scale VMD experiment for ammonia−water separation. The predicted permeation rate, ammonia recovery ratio, and ammonia concentration on the permeate side were in good agreement with the experimental results at different inlet flow rates on the feed side. The relative errors of the CFD model in the prediction of ammonia permeation rate, recovery ratio, and ammonia concentration on the permeate side were 8, 1, and 5%, respectively. A parametric study of the effect of the inlet ammonia concentration on the feed side and vacuum pressure applied to the permeate side was conducted to understand the flow and mass transfer characteristics, and the relationship with the module performance was discussed. Sensitivity analysis of the process parameters to the module performance was performed using the simulation results obtained from the parametric study.