Previous studies have indicated maternal exposure to particles with aerodynamic diameter <2.5 μm (PM2.5) is associated with preterm birth (PTB). However, no study has investigated this effect in pre-pregnancy impaired fasting glucose (IFG) women. This study aimed to differentiate the effects of maternal PM2.5 exposure on PTB between pre-pregnancy IFG and normoglycemia women, and to further identify the susceptible window. This cohort study was conducted between January 2014 and December 2017 in 21 Chinese cities. All the recruited women received pre-pregnancy fasting serum glucose (FSG) tests and were followed up for their delivery outcomes. The PM2.5 exposures were estimated by the daily air pollution concentrations of the nearby monitors. Women with FSG below 7.0 mmol/L were included in the analysis. We employed the Cox proportional hazards models to examine whether PM2.5 exposure was associated with PTB. 237957 women were included and 7055 (3.0%) of them were pre-pregnancy IFG. During the entire pregnancy, we found 24.1% (HR = 1.241; 95% CI: 1.069, 1.439), 61.8% (HR = 1.618; 95% CI: 1.311, 1.997) and 18.6% (HR = 1.186; 95% CI: 1.004, 1.402) of increases in risk for all PTB, early PTB (20–33 gestational weeks) and late PTB (34–36 gestational weeks) among the pre-pregnancy IFG women, and 15.9% (HR = 1.159; 95% CI: 1.127, 1.192), 33.9% (HR = 1.339; 95% CI: 1.255, 1.430) and 13.2% (HR = 1.132; 95% CI: 1.098, 1.168) of increases in risk for all PTB, early PTB and late PTB among the normoglycemia women, with each 10 μg/m3 increment of PM2.5 exposure, respectively. Furthermore, PM2.5 exposure had the strongest effect on all PTB during trimester 1 (0–12 gestational weeks) among the pre-pregnancy IFG women, compared with the less strong effect during trimester 1 among the normoglycemia women. In conclusion, pre-pregnancy IFG increases the risk of PTB attributed to PM2.5, especially during trimester 1. Moreover, the effects of PM2.5 are greater on early PTB than late PTB for both pre-pregnancy IFG and normoglycemia women.