BACKGROUND: To investigate the prevalence of obesity and diabetes among adult men and women in India consuming different types of vegetarian diets compared with those consuming non-vegetarian diets. METHODS: We used cross-sectional data of 156,317 adults aged 20-49 years who participated in India's third National Family Health Survey (2005-06). Association between types of vegetarian diet (vegan, lacto-vegetarian, lacto-ovo vegetarian, pesco-vegetarian, semi-vegetarian and non-vegetarian) and self-reported diabetes status and measured body mass index (BMI) were estimated using multivariable logistic regression adjusting for age, gender, education, household wealth, rural/urban residence, religion, caste, smoking, alcohol use, and television watching. RESULTS: Mean BMI was lowest in pesco-vegetarians (20.3 kg/m2) and vegans (20.5 kg/m2) and highest in lacto-ovo vegetarian (21.0 kg/m2) and lacto-vegetarian (21.2 kg/m2) diets. Prevalence of diabetes varied from 0.9% (95% CI: 0.8-1.1) in person consuming lacto-vegetarian, lacto-ovo vegetarian (95% CI:0.6-1.3) and semi-vegetarian (95% CI:0.7-1.1) diets and was highest in those persons consuming a pesco-vegetarian diet (1.4%; 95% CI:1.0-2.0). Consumption of a lacto- (OR:0.67;95% CI:0.58-0.76;p < 0.01), lacto-ovo (OR:0.70; 95% CI:0.51-0.96;p = 0.03) and semi-vegetarian (OR:0.77; 95% CI:0.60-0.98; p = 0.03) diet was associated with a lower likelihood of diabetes than a non-vegetarian diet in the adjusted analyses. CONCLUSIONS: In this large, nationally representative sample of Indian adults, lacto-, lacto-ovo and semi-vegetarian diets were associated with a lower likelihood of diabetes. These findings may assist in the development of interventions to address the growing burden of overweight/obesity and diabetes in Indian population. However, prospective studies with better measures of dietary intake and clinical measures of diabetes are needed to clarify this relationship.