<p>Accurate estimation of population density and spatial distribution is essential for wildlife conservation and management, yet conventional survey methods face substantial limitations. Estimating the density of medium- to large-sized mammals remains particularly challenging due to difficulties in direct observation and methodological constraints. This study qualitatively evaluated the detection performance and species identification capability of real-time drone surveys and assessed the applicability of distance sampling for density estimation. Surveys were conducted in a reedbed in Yeonsu, Incheon, and a plantation forest in Chungju, South Korea. Real-time drone surveys combined with line-transect sampling were applied in Yeonsu, whereas point-transect sampling was used in Chungju. Across nine line-transect surveys, 270 raccoon dogs (<i>Nyctereutes procyonoides</i>) were recorded, yielding an estimated density of 31.8 ind/km² with a relatively high detection probability (73.7%). In contrast, four point-transect surveys detected 35 water deer (<i>Hydropotes inermis</i>), resulting in an estimated density of 30.6 ind/km² but with low detection probability (34.4%) and wide confidence intervals due to limited sample size and complex terrain. These results indicate that real-time drone surveys, when integrated with distance sampling, can effectively correct for detection bias in density estimation of medium- to large-sized mammals. Future studies should apply advanced approaches such as Multiple Covariate Distance Sampling to account for variability in detection probability across survey conditions.</p>