Abstract： Objective:Explore the predictive value of body composition and related factors in early detection of gestational diabetes mellitus (GDM).Methods:949 pregnant women (142 cases in GDM group and 807 cases in normal group) in early pregnancy were selected from March 2019 to March 2020 in Suzhou Municipal Hospital, subject's clinical data were recorded. Body composition was measured by bioelectrical impedance method before the 13th week of pregnancy, and the relationship between age, protein, basal metabolic rate, body mass index (BMI) before pregnancy, body fat percentage, fat mass index (FMI), pregnancy times and the screening results of glucose tolerance in the second trimester of pregnancy were analyzed. The risk factors of GDM were further identified by multivariate regression analysis. Finally, the ROC curve was drawn to determine the diagnostic value of GDM, and the best boundary value was found to calculate the sensitivity and specificity of the indicators.Results:The age (20.82±2.60 vs 22.35±3.64), BMI before pregnancy (20.82±2.60 vs 22.35±3.64), percentage of body fat (29.37±5.63 vs 32.14±5.77), FMI [6.06(5.00, 7.30) vs 6.87(5.60, 8.60)] and pregnancy times [1(1, 2)vs 2(2, 3)] in GDM group were higher than those in normal group. Pregnancy times ( OR=1.232, 95% CI: 1.033-1.471) and FMI ( OR=1.228, 95% CI: 1.057-1.426) are independent risk factors of GDM. When FMI was used to predict the incidence of GDM, the area under the curve (AUC) was 63.0%. Conclusion:Pregnancy times and FMI in early pregnancy can be used as independent predictors of GDM. They provide a basis for scientific adjustment of diet and reasonable exercise, thereby preventing the GDM as early as possible. FMI can be reduced by adjusting the dietary structure and engaging in reasonable exercise, to reduce its risk among pregnant women.