Abstract: The present invention relates to a system for predicting gestational diabetes mellitus (GDM) of pregnant individuals, wherein the system comprises an operation module, and the operation model comprises a support vector regression model, and the system is used to predict the plasma glucose levels at 1 hour and/or 2 hours of oral glucose tolerance test (OGTT) by using a support vector regression developed prediction model generated by substituting the concentration of the biomarkers in fasting blood samples of pregnant individuals. The present invention provides biomarkers and biomarker-based diagnostic models for differential diagnosis of gestational diabetes mellitus (GDM), which can be applied to diagnosis or prediction of GDM in early stage and are of great significance to the prevention or treatment of GDM.