Knowledge of the Soil Water Retention Function (SWRF) is important for irrigation management, soil water modelling and quantification of soil phys- ical health. However, determining the SWRF is laborious, requiring spe- cialized skills and equipment. Multi-depth soil moisture data is routinely measured and collected by Australian farmers to guide irrigation man- agement. This data provides an extraordinary opportunity to better un- derstand and characterize the hydraulic properties of agricultural soils if a means of inversely determining the SWRF from daily multi-depth soil moisture data existed. While the SWRF has been solved using the soil water program HYDRUS-1D to inversely determine the van Genuchten- Mualem (VGM) parameters from inflow/outflow type experiments, its im- plementation in real-world, multi-layer soils is rare. This study evaluated the ability of the HYDRUS-1D model to determine the VGM parameters from daily soil moisture and weather data obtained at two Tasmanian field sites. The experiments were conducted as four separate 100-day periods from two sites for three soil depths, including five different optimized pa- rameter sets, and three different initial parameter value sources with each of six variants, producing 420 inverse modelling tests. The quality of these tests was assessed according to five criteria: (i) modelling robustness, (ii) simulation accuracy, (iii) prediction accuracy, (iv) parameter uniqueness, and (v) estimation stability. Overall, 322 of 420 simulations failed due to non-convergence. The accuracy of the successfully completed simulations was poor, with an average R ² value of 0.460 and an average RMSE value of 0.066 between the simulated and measured soil moisture. The prediction accuracy of the completed predictions of VGM parameters was less than 30%. Furthermore, predicted VGM parameters showed nonuniqueness and stability below 50%.