Compared to burn interiors, edges exhibit distinct biotic and abiotic conditions that include microclimate, wind speed, sunlight levels, soil composition, moisture content, nutrient availability, population density, and species diversity. This study characterized the landscapes in which burned forest edges formed in Samcheok, Korea. Over the study area, 500-m2 grid cells were generated to capture landscape characteristics. Grid cells intersecting burn boundary lines were designated as edge, while cells without these lines were classified as the interior of burned areas. Topographic variables including slope, elevation, topographic wetness index, solar radiation index, and proportions of fuel and land use types within grid cells were computed in a geographical information system(GIS). Correlation analysis with modified t-test and regression tree analysis were performed to explore the influences of landscape variables on edge formation with avoiding spatial autocorrelation problems. The results indicated that edges formed at low elevations with mild slopes, high topographic wetness, and low solar radiation. Edges were unlikely to form in areas dominated by Japanese red pines at low elevations. Moreover, heterogeneous land use/cover types contributed significantly to edge formation. Different forest management strategies for different landscape conditions can be more effective for enhancing resilience of forests to fire. Reducing susceptible fuel types might be effective at low elevations, while enhancing forest heterogeneity might be more effective at high elevations. Compared to burn interiors, edges exhibit distinct biotic and abiotic conditions that include microclimate, wind speed, sunlight levels, soil composition, moisture content, nutrient availability, population density, and species diversity. This study characterized the landscapes in which burned forest edges formed in Samcheok, Korea. Over the study area, 500-m2 grid cells were generated to capture landscape characteristics. Grid cells intersecting burn boundary lines were designated as edge, while cells without these lines were classified as the interior of burned areas. Topographic variables including slope , elevation, topographic wetness index, and proportions of fuel and land use types within grid cells were computed in a geographical information system (GIS). Correlation analysis with modified t-test and regression tree analysis were conducted to explore the influences of landscape variables on edge formation with avoiding spatial autocorrelation problems results indicated that edges formed at low elevations with mild slopes, high topographic wetness, and low solar radiation. Edges were unlikely to form in areas dominated by Japanese red pines at low elevations. Moreover, heterogeneous land use / cover types contributed significantly to edge formation Reducing susceptible fuel types might be effective at low elevations, while enhancing forest heterogeneity might be more effective at high elevations.