<p>In this paper, three-dimensional numerical modellings were carried out to thoroughly investigate the distribution and variation of reinforcement tensile force in the reinforced soil foundations (RSFs) overlying soft soil layer. The numerical models were established using the finite difference program FLAC^3D and were then calibrated and verified based on the measured results of field tests conducted on the geogrid-reinforced granular fill bed over soft clay. Furthermore, a parametric study was performed to ascertain the impact of various factors including the layout parameters and properties of reinforcement, the properties of soils and the thickness of backfill soil on the reinforcement tensile forces as the RSFs were at initial loading stage (s/B = 3%) and ultimate bearing capacity state (s/B = 10%). The numerical results indicate that at s/B = 3%, the distribution shapes of tensile forces were different for different layers, while at s/B = 10%, the tensile force distribution predominantly exhibited the saddle-shaped curves with the maximum tensile force (T_max) occurring near the footing edges for all layers and the T_max increased with the embedment depth of the reinforcements. In general, the T_max of tensile forces increased with the increase of the embedment depth of topmost layer and the vertical spacing between layers. However, the T_max firstly increased and then slightly decreased with the reinforcement length at s/B = 10%. Increasing the number of layers reduced the T_max in the same reinforcement layer. The tensile stiffness of reinforcement had significant effect on the distribution and magnitude of tensile forces. The larger tensile stiffness the larger tensile force. Compared with the elastic modulus and dilation angle of backfill soils, as well as the cohesion of soft soil, the friction angle had greatest influence on the tensile forces. At s/B = 10%, the T_max generally decreased first and then kept almost constant with an increase in the thickness of backfill soil between bottommost reinforcement layer and top of soft soil. The findings will be instrumental for engineers in the design and construction of geosynthetic-reinforced foundations over soft soil layer. Additionally, the results obtained can provide a basis for researchers to further perform similar numerical modellings and to validate the results of analytical solutions.</p>