Statistical experimental design, that is, response surface methodology, was used to predict and explain the effects of rubber ratio, carbon black, and accelerator level on the cure characteristics and physical properties of natural rubber/bromobutyl rubber (NR/BIIR) blends. With these three independent variables, 20 designed compounds were mixed by a two-roll mill and the scorch time, cure time, cure rate index, together with physical properties (hardness, tensile property, and compression set) were all determined by one operator. Multiple linear regression analysis was used to obtain response equations and thus contour plots, which illustrate the effects of the three independent variables on each property, as shown in detail by the diversity of interactions between independent factors and each property. It was found that the carbon black level is the most significant influential factor on scorch time, cure time, tensile properties, hardness, and compression set. The difference in reactivity toward sulfur vulcanization of NR and BIIR resulted in cure behavior and physical properties that are dominated by the NR content in the rubber ratio factor. Finally, the response equations were shown to be useful for making accurate predictions. (C) 2003 Wiley Periodicals, Inc.