Even though the optimum sampling methods for invasive pests are very important in newly invaded areas, the standard sampling unit of Ricania shantungensis is still undeveloped in persimmons. Among all developmental stages of R. shantungensis, the egg has close relationship between its density and subsequent tree damage. Thus, this study was conducted to suggest an optimum sampling unit for R. shantungensis eggs in persimmons based on characteristics of its within-tree distribution pattern. The within-tree distribution pattern was characterized with 60 persimmon trees by cutting 12 branches at three vertical levels (low, middle, and high) in four horizontal criteria (east, west, south, and north) per tree. The sampling units were determined based on coefficient of variation (CV) and coefficient of determination (r2) calculated from egg mass numbers per 10 cm from the tip within a branch. In numbers of R. shantungensis egg masses, there was no significant difference (P > 0.05) horizontally, but significant (P < 0.05) difference vertically. More R. shantungensis eggs were found on terminal branches of each trunk. The 60 cm from the tip of branches in the terminal positions of each trunk was selected as the optimum sampling unit for R. shantungensis in persimmons because this unit has the lowest CV value and more than 0.9 of r2 value. Even though the optimum sample number per tree should be determined field-specifically, it would be acceptable to sample two or three branches by considering this pests’ recognizable damage level. This small sampling unit could make the sampling of R. shantungensis become more economical, precise, and consistent in persimmon fields.