Background: Obese adults who are free from metabolic risk factors may develop risk factors over time. Our objective was to characterize development of obesity and duration of metabolically healthy obese (MHO) over 30 years. Methods: Participants in CARDIA who developed obesity (BMI ≥ 30 kg/m2) at follow-up exams during years 7, 10, 15, 20, 25, and 30 were analyzed. MHO was defined as obese and having 0 or 1 risk factor: ≥SBP/DBP 130/85 mmHg; fasting glucose ≥100 mg/dL/5.55 mmol/L; fasting triglycerides (≥150 mg/dL/1.69 mmol/L); and HDL-C (men <40 mg/dL/1.036 mmol/L, women <50 mg/dL/1.295 mmol/L) or on any medication(s) for these conditions. MHO duration (years) and obesity duration (years) were estimated for each subsequent time-point; and an overall cumulative duration was also calculated over available follow-up. MHO duration (%) was approximated as MHO duration ÷ obesity duration. Stable MHO was defined as 100% MHO duration over follow-up, while transient MHO was defined as <1–99%. Chi-squared tests were used to compare proportions by sex and race across obesity phenotypes. Multivariable-adjusted ANCOVA, adjusting for baseline BMI, age, race, and sex, was used to analyze obesity duration in all individuals who developed obesity, and also compare MHO duration (%) across race and sex in transient MHO individuals. Results: Of the 987 eligible participants who developed obesity, 51% were African American (AA), 56% were women. Higher percentages of AA were classified as transient MHO, and higher proportions of females were MHO (both p < 0.0001). Obesity duration (years) was higher in transient MHO compared with stable MHO (mean difference: 6.2 ± 0.5 years, p < 0.0001). Of those with transient MHO, African Americans (51.4 ± 1.6%) were more likely to have longer MHO duration compared to Caucasians (44.4 ± 1.9%, p = 0.005). Conclusion: MHO status can be a transient phenotype which differs by sex and race. Future studies are needed to explore modifiable lifestyle/behavioral predictors associated with longer MHO duration.