To derive an optimal illuminance of nighttime illumination, we conducted an experiment with 7 healthy young individuals and 7 healthy elderly individuals as subjects. After 20 minutes of adaptation to darkness, subjects were exposed to illumination under 5 conditions comprising 0.5 lx, 1 lx, 3 lx, 10 lx, or 30 lx vertical illuminance of the facial region, and heart rate variability (HRV) and electroencephalogram (EEG) were measured, and discomfort was evaluated by subjective report. Results of LF/(LF + HF) (LF = low frequency, HF = high frequency) demonstrated a V-shaped trend for the young groups beginning during exposure and ending post exposure, with 3 lx conditions representing the minimum value, a value markedly lower than that for 30 lx conditions. From these results we inferred that approximately 3 lx illuminance could best suppress physiological stress. Evaluation of discomfort by subjective report also demonstrated an increase in discomfort evaluation scores under high illuminance conditions. The alpha-wave proportion of EEG during exposure fell markedly on 3 lx or higher illuminance conditions, and we inferred that visual sensory information and cortical activity level were adequately attained in 3 lx or higher illuminance conditions. These results suggest that the optimal illuminance of nighttime illumination is approximately 3 lx.