This study aimed to assess the association between exposure to air pollution and unintentional injury deaths in South Korea. Data regarding all unintentional injury deaths (17,566) in seven metropolitan cities from 2002 to 2008 were collected. Using a time-stratified case-crossover study, conditional logistic regression and subgroup analyses were performed after stratification by age, gender, and season. To evaluate immediate and delayed effects of air pollutants, we used both single lag and distributed lag models. The risk was expressed as an odds ratio (OR) per one interquartile range (IQR) of each air pollutant. During the study period, the median (IQR) levels of air pollutants were 0.005 (0.004-0.007) ppm for sulfur dioxide (SO₂), 0.02 (0.02-0.03) ppm for nitrogen dioxide (NO₂), 0.03 (0.02-0.04) ppm for ozone (O₃), 48.3 (34.9-67.0) μg/m(3) for particulate matter ≤10 μm in aerodynamic diameter (PM₁₀), and 0.36 (0.1-0.6) ppm for carbon monoxide (CO). All air pollutants, with the exception of PM₁₀ and O₃, were significantly associated with an increased risk of unintentional injury deaths; the maximum risk was observed in the distributed lag 1 model for SO₂ (OR, 1.119; 95% confidence interval, 1.022-1.226), NO₂ (1.208; 1.043-1.400), and CO (1.012; 1.000-1.024). After stratification of the subjects by age, SO₂, NO₂, and CO were significantly associated with increased risk of unintentional injury deaths among subjects aged 60 years or older in the distributed lag 1 model, while O₃ and PM₁₀ were associated with increased risk among subjects aged 40 to 59 years. However, in subjects younger than 40 years of age, we found no significant associations for any of the air pollutants. Our study suggested evidence for a short-term association between air pollutants and unintentional injury deaths, even at low pollutants levels.