Wall-flow filters are worldwide recognized as the most efficient devices for the abatement of particulate emissions from automotive diesel engines. Mathematical models simulating the particulate thermal oxidation process in the filters are already applied for system optimization. This paper deals with the appropriateness of a specific assumption inherently used in all relevant published models, namely the temperature uniformity in the soot and wall layer. A new mathematical model is developed to predict the temperature gradients under various operating conditions. Based on the model results, it is shown that significant temperature gradients inside the soot layer may exist under some practical operating conditions. These conditions are associated with high flow rates and high soot loadings. In these cases, the uniform temperature assumption may lead to erroneous results for the prediction of the overall regeneration process. The error of this assumption is assessed as function of the soot porosity.