Accurately identifying hydrologically sensitive areas (HSAs) is crucial for effective watershed management. This study utilized digital soil mapping (DSM) technique to predict HSAs in the semi-arid watershed, Semnan province in central Iran. Four different pedo-transfer functions (PTFs) were employed to estimate saturated hydraulic conductivity (Ks). The Soil Topographic Index (STI) was calculated based on the soil properties in the study area. HSAs were identified using STI thresholds at three different levels: 9, 10, and 11. Additionally, the Soil and Water Assessment Tool (SWAT) model was utilized to analyze and validate the STI threshold method for defining HSAs. The random forest (RF) model performance in DSM, demonstrated coefficient of determination (R-2) values for soil organic matter (OM), bulk density (BD), clay, sand, and silt at soil depths of 0-30 cm and 30-60 cm, which were (0.78, 0.78, 0.60, 0.52, 0.71) and (0.66, 0.62, 0.69, 0.67, 0.49), respectively. Estimated Ks values ranged from 0.08 to 0.93 m day-1 (PTF1), 0.083-1.05 m day-1 (PTF2), and 0.039-1.25 m day-1 (PTF4), influenced by soil particle size distribution. HSAs under threshold values 9, 10, and 11 were approximately 35%, 16%, and 6% for PTF1; 35%, 18%, and 8% for PTF2; and 41%, 21%, and 10% for PTF4, indicating a decline in HSA area with increased STI intensity. Agricultural lands had the highest HSAs and runoff potential, linked to extensive human activity. This study shows that the HSA technique effectively identifies sensitive areas and potential runoff, with STI11 proving the most effective for prioritization across all PTFs. These findings highlight the importance of advanced modeling in watershed management for sustainable agriculture and improved water resource management.