We seek to observe the association between childhood obesity by different measures and adult obesity, metabolic syndrome (MetS), and diabetes. Thousand two hundred and nine subjects from "Beijing Blood Pressure Cohort Study" were followed 22.9 ± 0.5 years in average from childhood to adulthood. We defined childhood obesity using body mass index (BMI) or left subscapular skinfold (LSSF), and adult obesity as BMI ≥ 28 kg/m(2). MetS was defined according to the joint statement of International Diabetes Federation and American Heart Association with modified waist circumference (≥ 90/85 cm for men/women). Diabetes was defined as fasting plasma glucose ≥ 7.0 mmol/L or blood glucose 2 h after oral glucose tolerance test ≥ 11.1 mmol/L or currently using blood glucose-lowering agents. Multiple linear and logistic regression models were used to assess the association. The incidence of adult obesity was 13.4, 60.0, 48.3, and 65.1 % for children without obesity, having obesity by BMI only, by LSSF only, and by both, respectively. Compared to children without obesity, children obese by LSSF only or by both had higher risk of diabetes. After controlling for adult obesity, childhood obesity predicted independently long-term risks of diabetes (odds ratio 2.8, 95 % confidence interval 1.2-6.3) or abdominal obesity (2.7, 1.6-4.7) other than MetS as a whole (1.2, 0.6-2.4). Childhood obesity predicts long-term risk of adult diabetes, and the effect is independent of adult obesity. LSSF is better than BMI in predicting adult diabetes.