The experimental design and response surface methodology (RSM) is applied to a direct contact membrane distillation process. The factors considered for experimental design were the feed and permeate flow rate, the mean temperature, and the initial feed concentration of salt (NaCl) aqueous solution. The significant factors were optimized using a central composite design of orthogonal type. The quadratic models between the responses (permeate fluxes) and the independent parameters were built for both commercial and laboratory made membranes of different characteristics. The response surface models were tested with analysis of variance (ANOVA). For optimization purposes, the canonical analysis was employed. An algorithm was developed following the gradient method and using step adjustment in order to explore the response surface only inside the region of experimentation. The obtained optimal points were located in the valid region. The predicted permeate fluxes were compared with the experimental ones. In general, there is a good agreement between the experimental and the predicted permeate fluxes by RSM.