This paper is the theoretical study of the oxidative dehydrogenation of n-butane in porous membrane reactors. Performance of the membrane reactors was compared with that of conventional fixed-bed reactors. The porous membrane was employed to add oxygen to the reaction side in a controlled manner so that the reaction could take place evenly. Mathematical models for the fixed-bed reactor and the membrane reactor were developed considering non-isothermal condition and both radial heat and mass dispersion. From this study, it was found that the hot spot problem was pronounced particularly near the entrance of the conventional fixed-bed reactor. In addition, the assumption of plug flow condition did not adequately represent the reaction system. The effect of radial dispersion must be taken into account in the modelling. The use of the porous membrane to control the distribution of oxygen feed to the reaction side could significantly reduce the hot spot temperature. The results also showed that there were optimum feed ratios of air/n-butane for both the fixed-bed reactors and the membrane reactors. The membrane reactor outperformed the fixed-bed reactor at high values of the ratio. In addition, there was an optimum membrane reactor size. When the reactor size was smaller than the optimum value, the increased reactor size increased the reaction and heat generation and, consequently, the conversion and the selectivity to C-4 increased. However, when the reactor size was larger than the optimum value, oxygen could not reach the reactant near the stainless steel wall. It was consumed to react with the product, C4. As a result, the yield dropped. Finally, it was found that the increase of wall temperature increased the yield and that the feed air temperature could help control the temperature profile of the reaction bed along the reactor length.