【Objective】Urbanization absolutely needs activities like excavating, land-filling and road paving during its process, thus generating a variety of disturbed landform units, among which the dumping of waste soil is the main soil source supplying the construction of urban greenbelts with alien soil. Consequently, the distubered landform units join together with native landform units forming a complex underlying in the area of a urbanization construction project. 【Method】 In this paper, by means of field investigation and laboratory tests, a system comparison was made of the various landform units in material composition, water infiltration and water holding capacity, and then discussions were carried out on three types of reconstructed urban soil and their potential roles in mitigating the risk of urban water logging. 【Result】 Results show: (1) Soil bulk density differed significantly (p <0.05) between the landform units and varied in the order of Construction makeshift road (1.74 g cm^-3) > 2-year piles of dumped waste soil (1.58 g cm^-3) > 2-month piles of dumped waste soil (1.52 g cm^-3) > Waste grassland (1.47 g cm^-3)> Cultivated slopeing land (1.34 g cm^-3) > Artificial forest land (1.32 g cm^-3). In the piles of dumped waste soil, soil bulk density was higher on their flat tops than on their side slopes; (2) Soil infiltration rate varied in the order of side slopes of the piles of dumped waste soil > native landform > flat tops of the piles of dumped waste soil; stable infiltration rate on flat tops of the piles varied too with compactness of its surface layer, thus exposing it to the risk of forming of an aquitard; (3) Water storage capacity differed sharply from lanfform unit to landform unit (p<0.05), and was generally lower in reconstructed landform units (378.7 t hm^-2) than native ones (472.6 t hm^-2); it varied in the order of: 2-year piles > 2-month piles > Construction makeshift road, for disturbed landform units and in the order of Artificial forest land > Cultivated slopeland > Waste grassland for native landform units; and (4) For urban green land, soil reconstruction was usually done in three types, that is, tree-suitable, shrub-suitable and herb-suitable, and thickness bulk density, gravel content (<2 cm) and organic matter content were the main factors determining soil quality of the reconstructed soils. Soils reconstructed in line with the three types (herb-suitable, shrub-suitable and tree-suitable) may ensure the revegetation successful and capable of playing its potential role in regulating surface runoff and mitigating the risk of urban waterlogging, 2 months after plantation for herbs and 4~5 months for trees and shrubs. 【Conclusion】 All the findings in this study may provide certain scientific support for soil reconstruction for urban green belts and land and management of urban waterlogging during the process of urbanization in the Three-Gorge Reservoir Area.