Using pedotransfer functions (PTF) to estimate soil hydraulic properties may be necessary in soil water flow simulations for large-scale projects or in a pilot studies. The accuracy of a PTF outside of its development dataset is unknown. Existence of several models that are developed and tested in one region, but may perform relatively poorly in other regions, is common in meteorology which has developed the multimodel ensemble prediction techniques to address this problem. The objective of this work was to evaluate the estimation of soil water retention with a PTF ensemble. Data on soil water contents and pressure heads in 60 locations at 5 depths across a 6-m transect in layered loamy soil were collected during an extremely wet year in Belgium. Soil water fluxes were measured with passive capillary lysimeters at two depths. Water retention was measured in laboratory in samples taken at 60 locations at three depths. We used an ensemble of 22 PTFs developed from large datasets in different regions. The ensemble-estimated and laboratory water retention had comparable uncertainty (quantified as the width of 95% tolerance interval of water contents at a specific pressure head. The PTF ensemble gave a more accurate representation of field water retention compared with laboratory data. Monte Carlo simulations of soil water flow were done using the HYDRUS1D software with random sampling from laboratory and ensemble-estimated water retention data and with the saturated hydraulic conductivity estimated with a single PTF. Simulations with the ensemble-estimated water retention had, on average, two times smaller errors compared with laboratory water retention data. The same accuracy of simulating cumulative soil water fluxes was observed when either laboratory or ensemble-estimated water retention was used. The ensemble prediction of soil hydraulic properties is a promising method to estimate hydraulic properties in the uncertainty context.