This study aimed to neuroscientifically examine whether brain activity measured by functional near-infrared spectroscopy (fNIRS) during high-concentration fine dust education predicts actual mask-wearing behavior more accurately than self-reported mask-wearing intentions. Eighteen first-year middle school students viewed an approximately six-minute fine dust safety video while prefrontal cortex activation was recorded using a 48-channel fNIRS device. Immediately after viewing, students completed a self-reported mask-wearing intention survey (scored out of 100 on a Likert scale). Actual mask-wearing frequency was then tracked for two weeks. General linear model (GLM) analysis revealed that the aboveaverage mask-wearing group showed significantly greater activation in the left dorsolateral prefrontal cortex (DLPFC; BA 10) and left orbitofrontal cortex (OFC; BA 11) than the belowaverage group. Multiple regression analysis demonstrated that self-reported intention scores did not significantly predict actual behavior, whereas activation of the left OFC emerged as the only significant predictor of two-week mask-wearing frequency in this exploratory analysis. These findings suggest that neurophysiological responses to educational content may predict future health behavior more objectively and accurately than self-reports. This suggests that fNIRS-based measurement holds promise as a tool for evaluating the effectiveness of health and environmental education programs.