Response surface methodology (RSM) based on a five-level-three-variable central composite rotatable design (CCRD) was employed for optimization of preparation conditions of vinyltriethoxysilane (VTES)modified silicalite/polydimethylsiloxane (PDMS) hybrid pervaporation membranes. The three variables considered were silicalite loading, crosslinker/prepolymer weight ratio and polymer concentration. With a feed containing 5.0 wt% ethanol as a model solution, the main effects, quadratic effects and interactions of the three variables on the selectivity and total flux of hybrid membranes were investigated by the analysis of variance (ANOVA). The results showed that the main effect of silicalite loading was the most significant factor that influenced the hybrid membrane's selectivity, followed by the quadratic effect of silicalite loading, the main effect of crosslinker/prepolymer weight ratio and polymer concentration. The most significant factor that influenced the total flux was the main effect of polymer concentration. Regression equations between the preparation variables and the performance of the hybrid membranes were also established. Predicted values from the regression equations were found to be in good agreement with observed values, indicating that the regression equations could be used to predict and optimize the performance of the VTES modified silicalite/PDMS hybrid membranes. Under the preparation conditions of 66.8% silicalite loading, 0.098 crosslinker/prepolymer weight ratio and 23.7% polymer concentration, the maximum selectivity of 34.3 could be obtained with the feed containing 5.0 wt% ethanol at 323 K. (C) 2009 Elsevier B.V. All rights reserved.