Hydraulic conductivity (K) is a crucial parameter in hydrogeology but is highly heterogeneous and anisotropic due to variations in sediment texture, making its large-scale estimation challenging. Traditional laboratory and empirical methods based on grain-size distribution (GSD) analysis from limited data provide local K measurements, resulting in a poor representation of aquifer heterogeneity. In contrast, pumping tests estimate an integrated K value over a of the aquifer within the cone of depression but still lack the spatial resolution needed to reveal detailed variations in K across larger aquifer extents. In this study, the Di models method was used to simulate local GSD in three-dimensional (3-D) detrital systems. The focus was to explore the potential to estimate K through simulated particle-size fractions derived from a 3-D geological model of the City of Munich. By employing log-cubic interpolation, a complete and accurate representation of the fictive GSD enabled the application of multiple empirical relationships for K estimation. The resulting 3-D K fields preserved the variability in K within each aquifer system. When averaged for each separate aquifer system across different lateral extents, i.e., 50-150 and 550 m, the predicted K values showed success rates of 44-47% with deviations of at least one order of magnitude in 15-19% of cases when compared to 364 K values derived from pumping-test data. The results highlight the ability of the approach to successfully estimate K while accounting for spatial heterogeneity, suggesting its potential for groundwater modeling, aquifer yield assessments and groundwater heat pump system design.