The selection of a drag model is of critical importance for fluidized bed simulations. In this study, the effect of different drag models was investigated by conducting Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) simulations of bubbling fluidized beds and comparing the results with two sets of experimental data. For the data reported by Goldschmidt et al. (2004), the Di Felice model resulted in average particle height with less than 16% discrepancy, while the other drag models resulted in significantly lower values with discrepancies between 11 and 45%. For the NETL data (Gopalan et al., 2016), all the drag models showed reasonable qualitative agreement for the radial profiles of the solid velocities; however, no single model resulted in close quantitative predictions. None of the models were found to be suitable for both data sets. The analysis suggests that the Ayeni model and Di Felice model provide better predictions than the conventionally used Gidaspow model and Syamlal-O'Brien model. (C) 2018 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.