The 3D numerical simulations of gasliquid two-phase flows in three-anode and scaled-up 300 kA cell models using the computational fluid dynamicspopulation balance model (CFD-PBM) were presented. An Euler-Euler approach incorporated with a dispersed standard k-epsilon model and bubble-induced turbulence was used. The CFD-PBM was first validated with referenced experimental data. Then the effect of the electromagnetic force (EMF) was investigated. The basic flow patterns were predicted successfully by both the CFD and CFD-PBM simulations, but comparatively big differences remained in terms of gas bubble distributions. The bubble coalescence behavior played a predominant role under the anode bottom regions. The bubble breakup process became increasingly important once the different-sized bubbles escape from anode edge regions. Moreover, the overall distributions of various flow parameters in the simulations considering EMFs are clearly different from the cases without EMFs in an industrial 300 kA cell.