<p>The demand for groundwater resources for domestic, agricultural, industrial and commercial uses is increasing every year. Utilization of groundwater including spring sources is indispensable to address water shortages in the dry season. In this regard, spring source inventory mapping and identification of groundwater potential zones is urgently required in central Nepal. This study aims to identify groundwater potential zones in the hilly region of central Nepal using a logistic regression and random forest machine learning models. Ten groundwater influencing factors were considered: elevation, aspect, slope, curvature, drainage density, topographic wetness index, soil, land use, lithology and lineament density. A spring inventory map, comprising 161 spring locations was prepared through field surveys. The dataset was divided into training (80%) and test (20%) subsets, and a final groundwater potential map was created based on predicted probabilities. The accuracy assessment was evaluated using tenfold random cross-validation and receiver operating characteristic curves, which indicated that the area beneath the curve amounts to 0.82 for the logistic regression and 0.85 for the random forest method. The experimental findings illustrate the capability of these techniques to delineate groundwater potential zones in mountainous regions. These findings hold significant value for water management authorities. By leveraging this information, they can make informed decisions about sustainable groundwater use and future planning in the investigated area.</p>