Fluxes of the diffuse supernova neutrino background (DSNB) are calculated based on a new modeling of galactic chemical evolution, where a variable stellar initial mass function (IMF), depending on the galaxy type, is introduced and black hole (BH) formation from the failed supernova is considered for progenitors heavier than 18M ⊙. The flux calculations are performed for different combinations of the star formation rate, nuclear equation of state, and neutrino mass hierarchy, to examine the systematic effects from these factors. In any case, our new model predicts the enhanced DSNB flux at E ν ≳ 30 MeV and E ν ≲ 10 MeV, due to more frequent BH formation and a larger core-collapse rate at high redshifts in early-type galaxies, respectively. Event rate spectra of the DSNB at a detector from the new model are shown, and the detectability at water-based Cherenkov detectors, Super-Kamiokande with a gadolinium dissolution and Hyper-Kamiokande, is discussed. In order to investigate the impacts of the assumptions in the new model, we prepare alternative models, based on different IMF forms and treatments of BH formation, and estimate the discrimination capabilities between the new and alternative models at these detectors.