Soil erosion is a major environmental problem that threatens the sustainability and productivity of agricultural areas. Assessment and mapping of soil erosion are extremely important in the management and conservation of natural resources. This study was carried out to evaluate and compare the relationship between soil erodibility (K) maps prepared by traditional and geostatistical methods of the Sogulca Basin soils south of Ankara, Turkey. Ninety-three soil samples (0-20 cm) were collected from the study area to determine the soil erodibility (K). A kriged contour map was drawn based on the spatial variance structure of the data and was combined with the detailed soil map. A soil erodibility map formed using traditional methods underestimated the soil erodibility. A kriged K map displayed significantly better results than K map formed using traditional methods. The spatial variability of the K data changed with the land use and land form. The low erodibility class (K2) that was found 30.6% using a traditional approach decreased to 10.8% using a geostatistic method and the high erodibility class (K4) increased from 31.7% to 50.9%.