The discovery of dark energy from Type Ia supernovae (SN Ia) is based on the implicit assumption that the look-back time evolution of SN Ia luminosity, after light-curve corrections, would be negligible. A strong support for this assumption was the apparent insensitivity of SN Ia distances across the host galaxy morphologies. However, Hicken et al. 2009 (H09) shows a systematic difference in the Hubble residual (HR) of $0.144{pm}0.070$ mag between the E-S0 and Scd/Sd/Irr galaxies, after light-curve corrections. If true, this indicates that the light-curve fitters used by the SN Ia community can not correct for the population age (and therefore the evolution) effect. In order to confirm this, we have combined nearby SN Ia samples and the first-year SDSS-II SN Survey. The SNANA package was used for analyzing SN Ia light-curve, both for the MLCS2k2 and SALT2 fitters. We find a systematic difference in the HR of $0.10-0.13{pm}0.030$ mag between E-S0 and Scd/Sd/Irr galaxies, which is in agreement with the result of H09, but now at the 3-5 ${sigma}$ level. Considering the significant difference in the mean age of stellar population between these morphological types, the difference in the HR reported here suggests that the evolution effect of SN Ia luminosity should be considered in the cosmological application of SN Ia data.