There are several windows in the near-infrared wavelength range (~1 ¿m), in which thermal radiation emitted from the planetary surface penetrates through the thick Venus atmosphere and clouds. In this study we develop an improved method to estimate the surface emissivity on the basis of near-infrared thermal emission from the nightside Venus, which is observed through these windows. A simple radiative transfer model demonstrates that multiple reflection of thermal radiation between the atmosphere (including clouds) and the solid surface has a significant influence on the observed radiance under the condition of Venus, where reflectivity of overlying atmosphere and clouds is high. Thus it is necessary to take the effect of the reflection by the planetary surface into account in order to estimate accurately the variation in the surface emissivity on the basis of near-infrared observation of Venus. The net effect of multiple reflection of surface thermal radiation between the atmosphere and the surface is to significantly reduce the spatial contrast in thermal radiation due to surface compositional variation. The model calculation demonstrates that despite this effect, detection of granitic rocks on the Venus surface using near-infrared windows is feasible. Since granitic and basaltic rocks have dramatically different 1 ¿m emissivities, granitic rocks are distinguishable from basaltic rocks by ground-based telescopic observation. A Venus orbiter that measures both the near-infrared thermal radiation and the surface altitude with great accuracy will provide us with a reliable surface emissivity map of Venus, which is a very valuable tool to detect granitic (i.e., Earth-like continental) rocks on Venus.