Coke oxidation rates on supported metal catalysts were studied by temperature-programmed oxidation (TPO). The TPO data were successfully modeled with a linear combination of power-law kinetic expressions. The validity and limitations of the model, for carbon oxidation on metal containing catalysts, were analyzed in detail. The influence of heating rate, coke concentration, coke particle size and number, catalyst mass and oxygen concentration on TPO spectra predicted by the model agreed well with experimental results. Coke particle size and morphology have a great influence on a TPO spectrum. Even though this simplified form of kinetic expression does not explicitly account for all the physical phenomena in a coke-bum process, it considers some of the most important factors : the effect of carbon surface area and the promotional effect of supported metals on coke-bum kinetics. Deconvolution of a complex TPO spectrum by the above model provides kinetic parameters which can be used to predict TPO spectra of partial-coke-bum experiments. Difficulties in the predictive capability of this model arrive if other factors such as large variations in coke morphology and coke particle size exist and there is no information relating them to a particular feature in the TPO spectrum. However, the model can adequately account for these variations, if their characteristics and extent are known, e.g., they may be determined by electron microscopy.