The unit cell idealization has been long adopted in the settlement prediction of stone column-reinforced soils. This paper tests the accuracy of this modeling concept against trusted settlement values of engineering foundations. It is believed that in order to bestow the outcome of this study adequate generality different soil properties and foundation geometries need to be considered. It was, nevertheless, found impracticable to collect field settlement records for all the analyzed cases. The authors, therefore, appealed to the back analysis concept to construct a reliable mathematical model, calibrated against settlement records of full-scale field load test. This model, which is capable of reproducing the real field settlements, is then employed as a generic tool to obtain trusted settlement values for a variety of cases with essential geometrical similarity. The investigation revealed that the unit cell analysis may, in some cases, lead to erroneous estimation for the settlements of foundations with limited extents. Correction factors, dependent on the treated soil properties as well as the foundation size, are introduced.