A method for designing a fuel processor having an optimized size (i.e., volume and mass) for use in a fuel cell system which provides electrical power in a plurality of power ranges. The method includes maximizing water availability in the fuel cell system and sizing the first CO reduction reactor to provide for peak fuel cell system operational efficiency in a most-used power range while sizing the second CO reduction reactor to ensure the fuel processor can components to operate at a desired maximum power. The method allows development of a fuel processor that has significantly lower total mass and volume, and shorter start-up time, than conventionally designed processors, yet can perform at a desired maximum power.