Computational Fluid Dynamics (CFD) simulations of two-phase flow to predict dispersed phase hold up in a pulsed sieve-plate column have been performed and validated with reported experimental data. Suitability of different drag models for predicting dispersed phase hold up has been evaluated. Drag models applicable for concentrated dispersions are found to be better than the drag models applicable for lean dispersions. Kumar–Hartland drag model is found to be the best among the reported drag models with average error between predicted and experimental hold up being about 15%. Parameters of the Kumar–Hartland drag model are modified to further reduce the error to about 10%.