Abstract: Capillary barriers generally are considered for use in arid and semi-arid areas such that there have been few researches on capillary barrier covers in humid regions. A soil column was constructed in the open air to evaluate the performance of a capillary-barrier evaportranspiration cover. Percolation, surface runoff and water content variation under precipitation, evaporation and evapotranspiration were monitored for 18 months. A water balance model (HELP) and a soil-vegetation-atmosphere interaction model (VADOSE/W) were used to simulate the test. The total precipitation during the testing period is 2361 mm. There are 88.4 mm surface runoff and 67.4 mm percolation. Most precipitation is sent back to the atmosphere by the repeated water store-release process. The capillary barrier cover is effective because the hot season mostly coincides with the rainy season at the test site. A large amount of water is stored by the overlying clay layer during precipitation events due to the capillary break effect. Failure of the capillary break occurs only when the critical volumetric water content of the overlying finer textured clay layer is exceeded. The breakthrough is caused by consecutive intense rainfall events. Both HELP and VADOSE/W overestimate the surface runoff and percolation, and underestimate the evapotranspiration. The predictions given by VADOSE/W are more accurate than those given by HELP, because the behavior of unsaturated flows can be considered in VADOSE/W.