<p>This study presents a risk-based approach for mapping gully erosion susceptibility in the Jhang-Ping and Ping-Lin River watersheds of Taiwan. The study used six key conditioning indicators: rainfall, curve number, slope, topographic wetness index, distance to rivers, and distance to roads. Unlike the majority of recent studies that have relied on machine learning, this study used a simplified hazard–vulnerability framework, aiming to provide a transparent, interpretable, and cost-effective assessment tool. The model’s performance was validated on data from 21 field-surveyed gully sites. The model achieved an overall accuracy of 85.71%, a Kappa coefficient of 0.67, and an area under the receiver operating characteristic curve of 0.85. These results confirm the model’s robustness and practical applicability in terrain-based gully risk identification. Moreover, the proposed model offers straightforward management implications: high-risk zones (Levels 1 and 2) are recommended for engineering interventions, and lower-risk zones (Levels 3 and 4) may benefit from ecological stabilization measures. The proposed framework provides a valuable alternative to black-box algorithms, especially for regions with limited data availability or field resources, and contributes to bridging the gap between empirical erosion modeling and land management planning.</p>