A dynamic Monte Carlo (MC) simulation is presented for the catalytic reaction A + 1/2B(2) --> C. The catalyst surface is modeled as an array of crystallites each one represented as the top projection of a truncated hexagonal pyramid on a flat support. The proportion of corners, edges, base and face atoms or sites varies with crystallite size. The probabilities for the elementary steps comprising the catalytic cycle are calculated from the rate processes at which each elementary step occurs and varies in the different type of sites. We first analyze crystallites of the same size and then a catalyst with a non-uniform distribution of crystallites. The results show that the rate relative to a base case can increase or decrease with crystallite size depending on which elementary step is enhanced in the various types of sites analyzed. Results are relative to parameters used in the base case, which was taken from experimental results for the CO oxidation reaction. Monte Carlo simulations such as the one presented in this paper provide a method of analyzing the effect of crystallite size in the reaction, not available in continuos models. This method is today at the frontier of the modeling of chemical reacting systems and they will enhance our understanding of fundamental issues in catalytic reactions.