We performed cratering experiments on targets composed of glass beads with a power‐law size distribution that simulated the surface of rubble‐pile asteroids, and we improved the previously studied reduction factor on the crater size scaling relationship including the armoring effect using the momentum transfer efficiency. Cratering experiments were conducted using light gas guns at Kobe University and ISAS/JAXA to control the impact velocity, vi ${v}_{\mathrm{i}}$, from 50 to 4,400 m s−1. Two kinds of mixed targets were prepared by mixing glass beads with diameters of 0.1, 1, 3, and 10 mm—one with the smallest diameter beads, and one without. The size ratio of the target bead to the projectile (diameter of 1–3 mm), ϕ $\phi $, changed from 0.03 to 10. The crater size scaling relationships for the mixed targets were found to depend on the first contacted bead size. Notably, first contact with a 10 mm‐sized bead reduced the crater radius by 35% in maximum. The reduction factor due to this armoring effect on the crater size scaling relationship is written as follows: f(ϕ)=1+8.99×10−3vnormali0.630.34ϕ−0.19 $f(\phi )={\left[1+\left(8.99\times {10}^{-3}\right){v}_{\mathrm{i}}^{0.63}\right]}^{0.34}{\phi }^{-0.19}$; it decreased with the increase of the size ratio of the target bead to the projectile, while it increased with the increase of the impact velocity and approached unity. Our improved crater size scaling relationship that includes the reduction factor could be used to reconstruct the crater size frequency distribution on rubble‐pile asteroids such as Ryugu, and it may lead to a revision of the crater chronology of asteroids.