This theoretical study examined the impact of the degree of mixing, nonlinear particle breakage, and screen opening size on the particle size distribution (PSD) and mass hold-up in continuous dry mills with internal classification. A cell-based population balance model (PBM) incorporating a non-ideal screen model was formulated, wherein the back-mixing ratio and number of cells modulated the extent of axial mixing. The set of differential–algebraic equations (DAEs) was solved for the spatio-temporal evolution of the PSD in the mill and the product stream. The simulation results suggest that a smaller screen opening delayed the attainment of the steady state, increased the hold-up, and yielded a finer product PSD. The cushioning action of fines resulted in a coarser product PSD; however, a screen with a smaller opening mitigated this effect. The cell-based PBM predicted various features of experimental milling observations while providing insights into the mixing–nonlinear breakage–classification interplay.