Abstract Lanzhou is among the most seriously air-polluted cities in China as a whole, due to its unique topography, climate, industrial structure and so on. We studied the relationship between different air pollution and respiratory hospitalizations from 2001 to 2005, the total of respiratory hospital admissions were 28,057. The data were analyzed using Poisson regression models after controlling for the long time trend for air pollutants, the “day of week” effect and confounding meteorological factors. Three air pollutants (PM10, SO2, NO2) had a lag effect, the lag was 3–5 days for PM10, 1–3 days for SO2 and 1–4 days for NO2. The relative risks were calculated for increases in the inter-quartile range of the pollutants (139 μg/m3 in PM10, 61 μg/m3 in SO2 and 31 μg/m3 in NO2). Results showed that there were significant associations between air pollutants and respiratory hospital admissions, and stronger effects were observed for females and aged ≥65 yrs in Lanzhou.