For the reaction involving methanol oxidative dehydrogenation to formaldehyde: CH3OH 1/2O(2)/--> CH2O + H2O 1/2O(2)/--> CO + 2H(2)O, the performance of the packed-bed membrane reactor (PBMR) is compared with that of the conventional fixed-bed reactor (FBR) over a wide range of operating conditions. An experimentally validated reactor model is used for this purpose. It is found, both by simulations and experimental observations, that relative reactor performance depends strongly on the operating conditions. Using formaldehyde yield as the basis for optimization, optimal reactor performances for a fixed catalyst mass are determined and compared. The results predict higher optimal formaldehyde yield for the PBMR with oxygen fed via the membrane. The optimal reactor performance in this configuration is also less sensitive to variations in operating conditions and exhibits essentially 100% formaldehyde yield over a wide temperature range. (C) 2003 Elsevier Science Ltd. All rights reserved.