In this work, the performance of ceramic monoliths washcoated with Au/TiO2 is studied on CO preferential oxidation (CO-PrOx) reaction in H-2-rich environments under a wide range of operating conditions of practical interest. The parameter estimation of a nonlinear kinetic empirical model representing this system is made via genetic algorithms by fitting the model predictions against our laboratory observations. Parameter uncertainty leading to inaccurate predictions is often present when kinetic models with nonlinear rate equations are considered. Here, after the fitting was concluded, a statistical study was conducted to determine the accuracy of the parameter estimation. Activation energies of ca. 30 kJ/mol and 55 kJ/mol were adjusted for CO and H-2 oxidations, respectively. The catalyst showed appropriate activity and selectivity values on the CO oxidation on a H-2-rich environment. After ca. 45 h on stream the catalyst showed no deactivation. Results show that the model is suitable for reproducing the behavior of the CO-PrOx reactions and it can be used in the design of reactors for hydrogen purification. (C) 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.