Maternal exposure to ambient air pollution has been associated with preterm birth (PTB), however, entire pregnancy or trimester-specific associations were generally reported, which may not sufficiently identify windows of susceptibility. Using birth registry data from Guangzhou, a megacity of southern China (population ~14.5 million), including 469,975 singleton live births between January 2015 and July 2017, we assessed the association between weekly air pollution exposure and PTB in a retrospective cohort study. Daily average concentrations of PM, PM, NO, SO, and O from 11 monitoring stations were used to estimate district-specific exposures for each participant based on their district residency during pregnancy. Distributed lag models (DLMs) incorporating Cox proportional hazard models were applied to estimate the association between weekly maternal exposure to air pollutant and PTB risk (as a time-to-event outcome), after controlling for temperature, seasonality, and individual-level covariates. We also considered moderate PTB (32-36 gestational weeks) and very PTB (28-31 gestational weeks) as outcomes of interest. Hazard ratios (HRs) and 95% confidential intervals (95% CIs) were calculated for an interquartile range (IQR) increase in air pollutants during the study period. An IQR increase in PM exposure during the 20th to 28th gestational weeks (27.0 μg/m) was significantly associated with PTB risk, with the strongest effect in the 25th week (HR = 1.034, 95% CI:1.010-1.059). The significant exposure windows were the 19th-28th weeks for PM, the 18th-31st weeks for NO, and the 23rd-31st weeks for O, respectively. The strongest associations were observed in the 25th week for PM (IQR = 37.0 μg/m; HR = 1.048, 95% CI:1.034-1.062), the 26th week for NO (IQR = 29.0 μg/m; HR = 1.060, 95% CI:1.028-1.094), and in the 28th week for O (IQR = 90.0 μg/m; HR = 1.063, 95% CI:1.046-1.081). Similar patterns were observed for moderate PTB (32-36 gestational weeks) and very PTB (28-31 gestational weeks) for PM, PM, NO exposure, but the effects were greater for very PTB. We did not observe any association between pregnancy SO exposure and the risk of PTB. Our results suggest that middle to late pregnancy is the most susceptible air pollution exposure window for air pollution and PTB among women in Guangzhou, China.