Carbon-supported Pt-Ru alloy (Pt-Ru/C) catalysts were prepared using the "polygonal barrel-sputtering method". From the preparation of a Pt-Ru alloy with Pt/Ru ) ca. 50:50 atom % on a glass plate as support, the optimum sputtering conditions were an Ar gas pressure of 0.9-0.7 Pa and room temperature. The amount of the sputtered Pt-Ru alloy was controlled by changing the ac power and the sputtering time. Subsequently, the Pt-Ru/C samples were prepared under the given optimum conditions. The Pt-Ru alloy was dispersed extensively in the form of nanoparticles on a carbon support. For the ac power levels of 130, 100, and 50 W, the size distributions were narrower when the ac power was lowered. The respective average particle sizes were 4.1 nm (130 W), 3.3 nm (100 W), and 2.2 nm (50 W). In the case of 30 W, however, the size distribution and the average particle size were almost identical to those for 50 W. In addition, when the Pt-Ru/C samples were prepared by changing the sputtering time, only the dispersion density of the alloy nanoparticles increased in the Pt and the Ru deposited without changing the particle size. The atomic ratios of Pt and Ru in individual Pt-Ru alloy nanoparticles for the prepared samples were similar to the sputtering ratio and homogeneous compared with those for the commercially available samples. With regard to the electrochemical properties for the prepared samples, the hydrodynamic voltammograms for H 2 oxidation were identical to that of the commercially available sample. However, for CO oxidation, the peak shapes and the peak potentials for the prepared samples were sharper and ca. 20 mV lower than those for the commercially available samples, due to the uniform Pt and Ru atomic ratios of the individual alloy particles for the prepared samples. The coulomb charges of the CO oxidation reaction per amount of Pt and Ru for the prepared samples increased linearly in the reversed average particle sizes, while on the other hand, the charges for the commercially available samples were not proportional to the reversed sizes. This shows that the Pt-Ru alloy for the prepared samples was more efficiently utilized for electrochemical reactions rather than were the commercial ones. In addition, the cell performances for the alloy loading of 0.08 or 0.02 mg/cm 2 using the prepared Pt-Ru/C samples were similar to those for 0.50 mg/cm 2 using the commercially available sample.