This study aimed to determine whether increased carotenoids intake was associated with reduced risk of gestational diabetes mellitus (GDM). We performed a cross-sectional analysis using data from Tongji Maternal and Child Health Cohort study. The dietary carotenoids intake of 1978 pregnant women was assessed using a researcher-administered FFQ before undertaking an oral glucose tolerance test at 24–28 weeks. Multivariate logistic and linear regression analyses were used to obtain the effect estimates. Participants in the highest quartile of lycopene intake showed a lower risk of GDM (OR 0·50; 95 % CI 0·29, 0·86; Pfor trend = 0·007) compared with those in the lowest quartile; each 1 mg increase in lycopene consumption was associated with a 5 % (95 % CI 0·91, 0·99; Pfor trend = 0·020) decrease in GDM risk. No significant association was found between α-carotene, β-carotene, β-cryptoxanthin, lutein/zeaxanthin intake and GDM risk. Multiple linear regression analysis suggested an inverse association between lycopene intake and fasting blood glucose (FBG) (Pfor trend < 0·001); each 1 mg increase in lycopene intake was associated with 0·005 (95 % CI 0·002, 0·007; Pfor trend < 0·001) mmol/l decrease in FBG. Interaction analysis indicated consistent effect on each age or pre-BMI subgroup; however, a stronger protective effect of lycopene intake against GDM was observed among primigravid women (OR 0·20; 95 % CI 0·07, 0·55 in the highest v. the lowest quartile of intake; Pfor interaction = 0·036). In conclusion, dietary lycopene intake was mainly assumed via reducing FBG to decrease GDM risk, and the protection was relatively increased among primigravid women.