Abstract: Opencast mining has resulted in a large wasteland in the Jin-Shan-Mong region of western China. Local soils usually adapt to reconstruct the top soil layer in the existing reclamation strategy. The soil water-holding capacity can be improved to restore the land productivity. Among them, sandy loess and Pisha sandstone are widely distributed in the Jin-Shan-Mong region. But sandy loess has the lower soil water-holding, while Pisha sandstone has the lower porosity, which is limited as the reclamation topsoil. Therefore, the two- and three-layer reclamation strategies were proposed to increase the soil water-holding capacity in the Jin-Shan-Mong region. This study aims to investigate the effects of different reconstructed soils on the soil water dynamics during infiltration, drainage and evaporation. An indoor experiment of soil column was conducted on the reclamation materials with the Pisha sandstone, sandy loess and coarse sand. Five treatments of reconstructed soil were set as the 30 cm Pisha sandstone overlying 60-cm coarse sand (S1), 30 cm sandy loess overlying 60-cm coarse sand (S2), 15 cm Pisha sandstone and 15 cm sandy loess overlying 60 cm coarse sand (S3), 15 cm sandy loess and 15 cm Pisha sandstone overlying 60 cm coarse sand (S4), and 30 cm mixed soil overlying 60 cm coarse sand (S5). The optimal mixed soil was composed of 75% sandy loess and 25% Pisha sandstone. The infiltration test was carried out using a constant head method with a pressure of 5 cm at the tops of the soil columns. The soil columns were then completely saturated for drainage and evaporation. During drainage and evaporation, the total water storage was measured in each column by weighing, whereas the contents of soil water were measured at different depths using TDR probes. The hydraulic properties of four soils were measured in the laboratory and also optimized by inversion module in the Hydrus-1D. These optimized parameters were then used to simulate the infiltration, drainage and evaporation. The results showed that: 1) The reconstructed soil with Pisha sandstone as the top layer shared an outstanding retarding effect on the infiltration. The larger retarding effect was obtained with the increasing thickness of the Pisha sandstone layer, whereas, there was a smaller rate of stable infiltration when the soil column reached the steady seepage state. The cumulative time of infiltration was 1835, 168, 1095, 975, and 300 min, respectively, in the five treatments. 2) The total amount of evaporation was significantly higher in the S3 treatments than in the rest (P﹤0.05) .The smallest (32.1 mm) and largest (52 mm) amount of evaporation were found in the S4 and S3 treatments, respectively. The relative rate of evaporation was as low as 0.07 in the S4 treatment. 3) In the case of fine texture soil overlying coarse texture soil, the water loss primarily resulted from fine texture soil layer and the underlying coarse sand during evaporation. Additionally, the soil hydraulic parameters were optimized by the cumulative evaporation. The Hydrus-1D model was then used to simulate the infiltration, evaporation and drainage for all reconstructed soils. The larger water-holding capacity was achieved in the S5 treatment than the rest. Therefore, the S4 treatment was suitable for the soil reclamation in the Jin-Shan-Mong region.