A reaction-engineering model is presented, which describes catalyst performance as a function of the catalyst activity profile, the reaction kinetics, and the degree of catalyst deactivation. With this model, the catalyst activity profile can be optimised for Pt catalysed methyl alpha-D-glucopyranoside (slowly-reactive) and glucose (highly-reactive) oxidations. This is done by comparing modelling results with experimentally obtained data for catalysts of different activity distributions. Experiments in a semi-batch stirred reactor showed that for methyl a-Dglucopyranoside (MGP) oxidation at oxygen partial pressures below 40 kPa, egg shell catalytic activity distribution gives a higher rate of oxidation than a uniform distribution. It was also observed that with increase in oxygen concentration from 10 to 40 kPa, the rate of deactivation due to catalyst over-oxidation increased dramatically. For glucose oxidation, both catalyst activity distributions give the same oxidation rate for all investigated oxygen partial pressures (5-100 kPa). The developed model adequately describes the observed experimental results of both reactions. It was found that the active metal particle size has a significant influence on the catalyst deactivation for MGP oxidation; the uniform catalyst with higher dispersion shows a higher deactivation rate than the egg shell catalyst. For modelling glucose oxidation, the effect of catalyst particle-to-bubble adhesion and higher diffusivity or partition coefficient for oxygen have to be taken into account. (C) 2004 Elsevier B.V. All rights reserved.