The purpose of this study was to evaluate whether the apolipoprotein B/apolipoprotein A-I (apoB/apoA-I) ratio is a promising risk predictor of metabolic syndrome (MetS) and to determine the optimal cut-off value of this ratio in detecting subjects with MetS in a Chinese population. A prospective study was conducted using a representative sample of non-institutionized people in Taiwan. A total of 3,343 participants with mean age (±SD) of 39.86 (±15.61) years old were followed up from 2002 to 2007. The primary outcome was the incidence of MetS. The MetS was defined according to a unified criterion established by several major organizations. There were 462 cases of incident MetS during a mean follow-up period of 5.26 years. A significantly stepwise increase in the incidence of MetS across quartiles of the apoB/apoA-I ratio was noted in both sexes after adjustment for potential confounders (p for trend <0.001). Compared with the lowest quartile of apoB/apoA-I ratio, participants in the highest quartile had a significantly higher risk of MetS in both men [adjusted hazard ratio (HR) = 6.29, 95 % confidence interval (CI) = 2.79-9.13] and women (adjusted HR = 3.82, 95 % CI = 1.06-6.63). Comparisons of receiver operating characteristics curves indicated that the predictive ability of apoB/apoA-I ratio to detect MetS was better than conventional lipid ratio measurements. Furthermore, the optimal cut-off value of apoB/apoA-I ratio for MetS diagnosis was 0.71 in men and 0.56 in women. These results suggest that an elevated apoB/apoA-I ratio might constitute a potentially crucial measure linked to the risk of developing MetS.