This study presents results from a Kinetic Monte Carlo (KMC) simulation of the catalytic preferential oxidation of CO reaction (PROX) on 3D Pt crystallites. A new method of constructing normal distributions of event probabilities is presented based on the log transformation of calculated kinetic rates (log-KMC). This method allows for solving problems where the time scales of particular events are very different (stiff problems) and it was applied to the simulation of the PROX reaction on supported Pt catalysts to probe the experimental results on a catalyst with different crystallite sizes. It was found that the simulation replicated the trends from experimental results when it was assumed that the limiting step of the PROX reaction was oxygen adsorption. Lower Pt dispersions gave place to catalysts showing higher turnover frequencies, or TOF, which is due to large crystallites having lower numbers of less active base sites. This limiting effect is simulated in larger crystallite sizes with more available sites on (1. 1. 1) faces.