This study evaluated the physiological remineralization of artificially demineralized dentin beneath glass ionomer cements with and without bacterial contamination. The artificially demineralized dentin was produced on 84 monkey teeth using a decalcifying solution. Half the samples were left open to the oral cavity for one week, then, all cavities were restored with two glass-ionomer cements: Fuji IX or Fuji II LC improved (n = 7). The nanohardness of the artificially demineralized dentin at 3, 90 and 360 days was measured using a nanoindentation tester (ENT-1100, Elionix) and compared statistically by two-way ANOVA and Fisher's PLSD test (p < 0.05). Each mineral (Ca, Mg, P, F) within the demineralized dentin was also analyzed using Electron Probe Microanalysis. For the samples, the mean nanohardness of the three-day samples was significantly lower than the 360-day samples (p < 0.05). Although there was no significant difference in the mean nanohardness within all the bacterially-contaminated groups through the experimental periods (p > 0.05), the mean nanohardness of the bacterial-contaminated samples were significantly lower than the non-bacteria-contaminated samples (p < 0.05). From the EPMA results, fluoride release from both cements to the bottom of the artificially demineralized dentin was detected within three days. Although Ca density was sparse within this demineralized dentin lesion, for the Fuji IX sample, a high Mg density within this lesion was detected at 360 days.