This thesis report presents an investigation of a soil improvement technique that is being executed for a Van Oord project in Kuwait. The soil is improved by the use of a column supported embankment, consisting of sand columns installed in a soft soil layer and a sand platform. The efficiency of this method is defined in terms of stress transfer and settlement reduction. When the soil improvement is finished and the land will be used, there are conditions concerning bearing capacity and settlement behavior. To this extent two important parameters were defined. i.e. the incremental efficiency (the load increase in a sand column over the total surface load increase) and incremental settlement reduction ratio (the settlement of the improved soil over the settlement of the unimproved soil (i.e. soil that has not been improved by sand columns), under loading). To determine the efficiency of the soil improvement, a number of tests were performed on site. Tests included plate load tests (in this thesis referred to as zone load tests). The load tests were simulated in Plaxis, with the known load/settlement results the model could be benchmarked. Furthermore soil samples were taken and tested to determine the local soil characteristics. The parameters derived from the soil tests are also used in the Plaxis calculations. Plaxis allows for a step-by-step consolidation of the soft soil in which the columns were installed. It can be seen that the stress distribution changes for different stages of consolidation. The columns are first constrained by the very stiff soft soil layer (due to high excess pore pressures under loading). When the pore pressures dissipate the constraining stress is lowered and the column head expands. Under vertical loading the stress in the column head has a funnel shape, due to the displacements in the outer ring of the column head. Based on the Plaxis calculations it can be concluded that when a load is activated on top of a surface of soil that has been improved by the use of sand columns (with a center-to-center distance of three meters), given that the platform is thick enough, 60% of that load is transferred to the column. With a greater center-to-center distance between the columns that percentage decreases, e.g. 28% for a column spacing of five meters. Compared to existing theories by Hewlett and Randolph (1988) and Zaeske (2001) (it should be noted that most existing theories assume presence of geosynthetic reinforcement, which is not the case for this project) the calculated column force is relatively low. A minimal thickness of the sand platform is needed to facilitate maximum efficiency. The thickness as determined by the Plaxis calculations are lower compared to existing literature. With platform heights of up to seven meters no full arching was observed, however partial arching did occur as evidenced by the efficiency values.