Full spectrum fitting is a powerful tool for estimating the stellar populations of galaxies, but the fitting results are often significantly influenced by internal dust attenuation. For understanding howthe choice of the internal dust correction method affects the detailed stellar populations estimated from the full spectrum fitting, we analyze the Sydney-Australian Astronomical Observatory Multiobject Integral field spectrograph (SAMI) galaxy survey data using the Penalized PiXel-Fitting (PPXF) package. Three choices are compared: (Choice-1) using the PPXF reddening option, (Choice-2) using the multiplicative Legendre polynomial, and (Choice-3) using none of them (no dust correction). In any case, the total mean stellar populations show reasonable mass-age and mass-metallicity relations (MTR and MZR), although the correlations appear to be strongest for Choice-1 (MTR) and Choice-2 (MZR). Whenwe compare the age-divided mean stellar populations, theMZRof young (<109.5 yr ≈ 3.2 Gyr) stellar components in Choice-2 is consistent with the gas-phase MZR, whereas those in the other two choices hardly are. On the other hand, the MTR of old (≥109.5 yr) stellar components in Choice-1 seems to be more reasonable than that in Choice-2, because the old stellar components in low-mass galaxies tend to be relatively younger than those in massive galaxies. Based on the results, we provide empirical guidelines for choosing the optimal options for dust correction.