Based on a stochastic mesoscopic model, the influence of internal noise on the oscillatory kinetics of the catalytic oxidation of CO on nm-sized palladium particles is studied, using the chemical Langevin equations, Poisson approximation algorithm, and exact stochastic simulation algorithm. The reaction rate oscillations are of stochastic nature due to considerable internal noise in such mesoscopic systems. It is found that the performance of the stochastic oscillations undergoes a maximum with the variation of internal noise level for a given CO partial pressure, which demonstrates the occurrence of internal noise stochastic resonance. This phenomenon implies that optimal internal noise would favor the reaction rate oscillation of CO oxidation on nm particles. Such a phenomenon is robust to the change of external parameters, such as CO pressures.