Objective: To understand the association between gestational weight gain (GWG) per trimester/total (GWG) and gestational diabetes mellitus (GDM). Methods: A prospective cohort study was conducted among 829 healthy pregnant women from Chengdu who delivered a singleton during March-September 2013. Basic information about pre-pregnancy was collected at the first time, exercise information per trimester, dietary intake information per trimester and gestational age at delivery were obtained by questionnaire. Weight at the (12 ± 1)th, (28 ± 1)th, (36 ± 1)th and last week before delivery were measured. GDM was diagnosed according to the National Diabetes and Pregnancy Prevention Guidelines (2014). At last, a total of 682 pregnant women were analyzed. After controlling the potential confounders, multiple logistic regression models were used to test the associations between GWG per trimester/total GWG and GDM. Results: After adjusting for age at delivery, pre-pregnancy body mass index, family histories of type 2 diabetes and hypertension, exercise in the first trimester, parity and energy intake in trimester and other potential confounders. Compared with adequate GWG in the first trimester, insufficient and excessive GWG in the first trimester were associated with increased risk of GDM (OR=1.23, 95%CI: 0.63-2.38, and OR=2.20, 95%CI: 1.12-4.35). Compared with adequate GWG in the second trimester, insufficient and excessive GWG in the second trimester were associated with decreased risk of GDM (OR=0.47, 95%CI: 0.18-1.19, and OR=0.78, 95% CI: 0.43-1.42). Compared with adequate GWG in the third trimester, insufficient GWG in the third trimester was associated with increased risk of GDM (OR=1.48, 95%CI: 0.77-2.84), excessive GWG in the third trimester was associated with decreased risk of GDM (OR= 0.53, 95% CI: 0.28-0.99). Compared with adequate total GWG in pregnancy, insufficient GWG in pregnancy was associated with increased risk of GDM (OR=2.16, 95% CI: 1.04-4.46), excessive GWG in pregnancy was associated with decreased risk of GDM (OR=0.74, 95% CI: 0.38-1.46). Conclusions: Insufficient and excessive GWG in the first trimester were associated with increased risk of GDM, the first trimester may represent a critical period for risk of GDM. The effects of GWG in the second and third trimester on GDM need further study.