While the importance of spatial scale in ecology is well established, few studies have investigated the impact of data grain on hotpot analysis of biological resource. In this study, in order to analyze the impact of spatial scale on hotspot analysis of Antarctic krill (Euphausia superba) density distribution in the Southern Ocean, the krill resource density data in 1926 to 2016 by 10-year period was interpolated into ten spatial scales, i.e., 10′×10′, 20′×20′, 30′×30′, 40′×40′, 50′×50′, 1°×1°, 2°×2°, 3°×3°, 4°×4°, and 5°×5°, respectively, and linear, logarithmic, exponential, power-law, and polynomial functions were used to calculate the relationship between the global krill resource density, krill resource density in hotspot (coldspot) areas and the spatial scale, and the variations in centroid and hotspot (coldspot) areas at different spatial scales were also analyzed. The results showed that there was a significant scaling relationship between the maximum, skewness, kurtosis, Coefficient of Variation (CV) of global krill resource density and the spatial scale. Significant scaling relationship between the maximum, skewness, kurtosis, third quartile (Q3), and CV of the krill resource density in hotspot area and the spatial scale was found. A significant scaling relationship were also found between the maximum, mean, standard deviance (SD), skewness, kurtosis, Q3, CV and spatial scale for the coldspot area. The size of the hotspot (coldspot) area increased with the increasing of the spatial scale, and the offset of centroid indicated the locations of the hotspot (coldspot) area were greatly affected by the spatial scale. Significant shift of centroid for hotspot (coldspot) areas occurred when the spatial scale was greater than 1°×1°. Therefore, it is recommended to use spatial scale that less than 1° × 1° to identify the local spatial pattern for hotspot analysis of krill resource density.