As we demonstrated in Paper I, the quenched fractions of central and satellite galaxies as function of halo mass are extremely similar, as long as one controls for stellar mass. The same holds for the quenched fractions as a function of central velocity dispersion, which is tightly correlated with black hole mass, as long as one controls for both stellar and halo mass. Here we use mock galaxy catalogs constructed from the latest semi-analytic model, L-GALAXIES, and the state-of-the-art hydrodynamical simulation, EAGLE, to investigate whether these models can reproduce the trends seen in the data. We also check how the group finder used to identify centrals and satellites impacts our results. We find that L-GALAXIES fails to reproduce the trends. The predicted quenched fraction of central galaxies increases sharply with halo mass around $10^{12.5}h^{-1}M_{odot}$ and with black hole mass around $sim10^{6.5}M_{odot}$, while the predicted quenched fraction of satellites increases with both halo and black hole masses gradually. In contrast, centrals and satellites in EAGLE follow almost the same trend as seen in the data. We discuss the implications of our results for how feedback processes regulate galaxy quenching.