Jakarta has long promoted infiltration wells as a key component of urban stormwater management, yet their performance is often limited by fine-grained alluvial soils and spatially variable groundwater depths. This study presents a field-informed assessment of modified infiltration wells for hydrogeologically constrained urban settings, using Jakarta as a representative deltaic case. Geological mapping, groundwater-depth mapping, in-situ infiltration testing, hand-bored sampling, laboratory falling-head permeability testing, and unsaturated hydraulic characterization were integrated with transient seepage modeling in SEEP/W to evaluate both conventional and modified infiltration-well configurations. Field investigations showed that hydraulic conductivity commonly decreases to less than 1 × 10⁻⁸ m/s below shallow depth, suggesting that conventional infiltration wells may have limited effectiveness in many parts of the city. Under original soil conditions, the simulated fraction of rainfall transferred to the well was only 0.56%. A modified configuration incorporating an engineered sand replacement layer and, where required, an HDPE geomembrane appeared to improve performance substantially, increasing effectiveness to 53.42% under shallow-groundwater conditions and to 73.28% where a deeper groundwater table provided greater unsaturated storage. The results suggest that the feasibility of infiltration-based stormwater systems in Jakarta depends on the combined effects of subsurface permeability and available unsaturated thickness above the groundwater table. Rather than proposing a new physical mechanism, the study provides practical design guidance for adapting infiltration wells to restrictive urban subsurface conditions and supports the use of groundwater-sensitive design criteria in place of uniform area-based rules. © 2026 The Author(s)