The purpose of the present study was to assess the risk of the metabolic syndrome (MS) with vegan, pescovegetarian, lactovegetarian and nonvegetarian diets in Taiwan. The design was a retrospective cohort study using secondary data analysis from a Taiwan longitudinal health check-up database provided by MJ Health Screening Center during 1996-2006. A total of 93209 participants were classified as vegans (n=1116), pescovegetarians (n=2461), lactovegetarians (n=4313) and nonvegetarians (n=85319) by food frequency list of self-administered questionnaire at baseline. The association between MS or MS components and different dietary groups was evaluated using Cox proportional-hazards regression models with adjustment for confounders. During the mean 3.75 years of follow up, a total 8006 MS incident cases occurred and the incidence of MS was 229 (95% CI, 224, 234) per 10000 person year. Compared with vegans, hazard ratios of MS for nonvegetarians, pescovegetarians, lactovegetarians were 0.75 (95% CI, 0.64, 0.88), 0.68 (95% CI, 0.55, 0.83) and 0.81 (95% CI, 0.67, 0.97) after adjusting for sex, age, education status, smoking status, drinking status, physical activity at work and leisure, respectively. As for MS components, nonvegetarians and pescovegetarians had 0.72 (95% CI, 0.62, 0.84), 0.70 (95% CI, 0.57, 0.84) times risk of developing low high density lipoprotein cholesterol (HDL-C), while nonvegetarians had 1.16 (95% CI, 1.02, 1.32) times risk of developing high fasting plasma glucose. Our data suggest that the vegan diets did not decrease the risk of metabolic syndrome compared with pescovegetarian, lactovegetarian and nonvegetarian diets in a Taiwanese cohort.