Due to computational time limitations, fully resolved simulations using the two-fluid model of the flow inside industrial-scale fluidized beds are unaffordable. The filtered approach is used to account for the effect of small unresolved scales on the large resolved scales computed with coarse realistic meshes. Using a fully resolved simulation, we highlight the need to account for a subgrid drift velocity to obtain the correct bed expansion when using coarse meshes. This velocity, defined as the difference between the filtered gas velocity seen by the particle phase and the resolved filtered gas velocity, modify the effective relative velocity appearing in the drag law. We close it as a correction of the resolved relative velocity depending on the filtered particle concentration and the filter size. A dynamic procedure is used to adjust a tuning parameter. Bed expansion obtained with a posteriori test on coarse-grid simulations matches well to fully resolved simulations. (C) 2011 American Institute of Chemical Engineers AIChE J, 2012