The estimation of subsurface drain spacing at large scales remains a challenging task due to the combined influence of heterogeneous land uses, complex topography, diverse soil types, and variable soil hydraulic properties. In this study, a combined approach is developed based on the SWAT (Soil and Water Assessment Tool) hydrological model and the well-established Donnan–Hooghoudt equation to assess subsurface drain spacing at the watershed level. The SWAT model was first employed for watershed delineation and subdivision into hydrologically meaningful sub-basins, accounting for spatial variability in land use, topography, soil characteristics, and climatic conditions. Open access datasets were used for the SWAT input geospatial layers and also to derive soil hydraulic parameters (e.g., saturated hydraulic conductivity) as well as crop distribution at the parcel level. For each sub-basin, key variables were determined, including dominant crop type, representative root depth, depth to the impermeable layer, and target groundwater table depth. Subsequently, the Donnan–Hooghoudt equation was applied within each sub-basin to estimate the appropriate drain spacing, tailored to its specific soil, crop, and hydrological requirements. The methodology was tested in the Zazari–Chimaditida sub-basin in Western Macedonia, Greece. The results demonstrate substantial spatial variability in drain spacing across the sub-basins, ranging from 26 m to 73 m. This variability highlights the critical role of SWAT-based watershed discretization and the spatial resolution of input datasets in drainage design.