Abstract: Objective To explore the association of dust events with daily respiratory hospitalization.Methods After controlling long time trend,calendar effect and other meteorological confouding factors,a semi-parametric generalized additive model(GAM)considered simultaneously the effects of dust events up to lags of 7 days was fit to explore the relationship between dust events and daily respiratory hospitalization from 1994 to 2003.At the same time,the association for different sex groups was assessed.Results (1)There were 3-day delayed effect of dust events on total daily respiratory hospitalization and upper respiratory tract infcetion(URTI),their relative risk(RR)were 1.14(95% CI=1.02～1.27)and 1.16(95% CI=1.00～1.36)respectively.(2)There were 3-day delayed effect on both male and female of total daily respiratory hospitalization,their RR were1.14(95% CI=1.01～1.29)and 1.18(95% CI=1.00～1.41)respectively,but only 4-day delayed effect on male of U R TI and 6-day delayed effect on male of pneumonia were found,their RR were 1.28(95% CI=1.04～1.59)and 1.17(95% CI=1.00～1.38)respectively.Conclusion There are the delayed effects of dust events on various daily respiratory hospitalizations;moreover,for URTI and pneumonia,there are close associations of dust events with male hospitalizations than females.