This paper attempts to provide one of the first population-based causal estimates of the effect of air pollution on suicidal ideation—a key precursor to suicide attempt and completion—among school-age children. We use daily variations in the local wind direction as instruments to address endogeneity in pollution exposure. Matching a unique risk behavior survey of 55,000 students from 273 schools with comprehensive data on air pollutants and weather conditions according to the exact date and location of schooling, our findings indicate that a 1% decline in daily PM2.5 is associated with a 0.36% reduction in the probability of suicidal ideation. Moreover, the dose-response relationship reveals that the marginal effects increase significantly and non-linearly with elevated concentration of PM2.5. The effect is particularly pronounced among younger, male, students from low-educated families, and students with lower grades.