We conducted a validation study of a newly developed near-surface soil moisture assimilation scheme for estimating effective soil hydraulic properties using soil moisture data from different hydroclimatic regions, including semihumid Oklahoma, humid Iowa and Illinois, and temperate humid China. A genetic algorithm (GA) was used to estimate the effective soil water retention θ(h) and hydraulic conductivity K(h) functions of an effective modeling domain by minimizing errors between observed near-surface soil moisture and values simulated with the Richards-based Soil–Water–Atmosphere–Plant (SWAP) model. The parameter estimation approach considered uncertainties in the initial and bottom boundary conditions, rooting depth, and root density by creating simulation ensembles based on combinations of several modeling conditions and a multipopulation approach in the GA to estimate uncertainties in the derived soil hydraulic properties. The results showed that θ(h) is not very sensitive to variations in the initial and boundary conditions, rooting depth, and root density applied on the modeling domain. The value of K(h) was found to be more sensitive to variations in rooting depth and root density than to variations in the initial and boundary conditions. With the modeling domain better represented, the estimated θ(h) and K(h) functions were found to be satisfactory in most of the locations studied. They were validated using laboratory-measured θ(h) and Ksat, observed soil moisture in the field, and soil hydraulic properties from the UNSODA database. Our study indicates, however, that the homogeneous-medium assumption commonly used to effectively describe a heterogeneous system may fail to closely represent a highly heterogeneous (layered) soil profile if only the near-surface soil moisture data are used to define the subsurface soil hydraulic properties. Additional soil moisture data from deeper depths may be needed to better estimate the effective soil hydraulic properties of highly heterogeneous systems.