This work focuses on the recovery of two odorous compounds (diacetyl and S-methylthiobutanoate) from dilute aqueous solutions by pervaporation using polyether block amide 2533 (PEBA) membrane. The pervaporation process was experimentally investigated combined with model simulation. The PEBA 2533 membrane was prepared by solvent casting method, and effect of temperature and concentration on its flux and permselectivity were studied through pervaporation experiments using binary model solutions. A correlation between experimental results and those obtained via numerical model as function of feed temperature and concentration using solution-diffusion theory describing flux and separation factor was established. The sorption properties of PEBA in the organic aqueous solution were predicted by the Flory-Huggins model and the interaction parameters were obtained from the Hansen solubility parameter method using group contribution method. A semi-empirical relationship of the diffusion coefficient for binary mixture through PEBA membrane was proposed as a function of concentration and temperature, which was validated by comparison with experimental data. The results showed that at a given temperature, the organic flux increased with increasing feed concentration while the separation factors decreased for the both the mixtures, while elevated temperature increased the total flux without sacrificing the separation factor. The maximum organic flux and separation factors for diacetyl/water and S-methylthiobutanoate/water mixtures, were 1.6 g center dot m(-2)h(-1) and 40, and 4.6 g center dot m(-2)center dot h(-1) and 110, respectively. The pervaporation performance for the binary mixtures system was believed to be dependent on diffusion process where PEBA membrane exhibited appreciable selectivity to S-methylthiobutanoate at low concentrations. The present approach based on the prediction of data using Flory-Huggins model and experimental results for the separation and extraction of S-methylthiobutanoate and diacetyl from aqueous solution could be of great help on industrial level applications and process designing.