In this work the application of genetic algorithms (GA) for solving the parameter estimation problem in a nonlinear empirical kinetic model for CO preferential oxidation is reported. Kinetic models with nonlinear rate equations often suffer of considerable parameter uncertainty which can lead to inaccurate predictions. Here, after the parameter vector is obtained, a statistical study is performed in order to show how accurate the parameter estimations are determined. The unknown model parameters are obtained by fitting the model predictions against our laboratory observations measured under a range of experimental conditions using a novel Au/TiO<SUB>2</SUB> catalyst.Sociedad Argentina de Informática e Investigación Operativa (SADIO)