Objective: Increased obesity in the development of type 2 diabetes (T2DM) has not been thoroughly evaluated in prospective manner in women with previous history of gestational diabetes (GDM). In this study, we investigated the effect of change in BMI on the development of T2DM in women with previous history of GDM. Methods: Subjects with previous history of GDM were recruited and prospectively followed from four tertiary hospitals in Korea. The participants performed 75 g oral glucose tolerance test (OGTT) serially starting at 6 week and annually after the postpartum period. Change in BMI was calculated between the initial and the last visit of examination or at the onset of diabetes. For statistical analyses, subjects were stratified into the 3 groups based on the tertiles of BMI changes, and association with the onset of T2DM was examined. Results: 418 subjects were included for analysis. Mean duration of follow up was 3.9 ± 1.8 years. The BMI change in each tertile was -1.8 ± 1.1, -0.2 ± 0.3, and 1.6 ± 1.2 kg/m², respectively. Incident diabetes was observed in 53 subjects (12.7%). Incident diabetes were found in 8.6%, 12.6%, and 16.9% in each tertile group (P = 0.039). Worsening of glucose tolerance status according to BMI change was observed in 21.6%, 25.2%, and 39.7% (P < 0.001). Changes in clinical (e.g. systolic and diastolic blood pressure) and biochemical parameters (e.g. fasting glucose, insulin, and lipid profiles) were significantly improved in the lowest tertile. When change in BMI was evaluated as a continuous variable, it was an independent predictor of incident diabetes even after adjusting for age and duration of follow up (P = 0.025). Conclusion: This study demonstrates that postpartum BMI change during four years of follow up has a significant impact on the onset of diabetes. Our study implies that tight weight management after delivery in GDM population should be beneficial for the prevention of diabetes.