Epidemiological studies suggest that elevated circulating ferritin is associated with heightened incident diabetes in mainly Western populations, although the results were not entirely consistent. We aimed to prospectively investigate the ferritin-diabetes association in an Asian population for the first time, to our knowledge, and also to examine this association with an updated meta-analysis. Our prospective study included 2198 community-living Chinese between 50 and 70 y of age in 2005. All individuals participated in a 6-y follow-up survey in 2011. Fasting plasma ferritin, high-sensitivity C-reactive protein (hsCRP), adiponectin, and γ-glutamyltransferase (GGT) were measured at baseline. A total of 538 incident diabetes cases were documented by self-reports and/or fasting glucose ≥7.0 mmol/L at the follow-up survey. After multiple adjustments, the RR of type 2 diabetes was 1.90 (95% CI: 1.37, 2.65) when comparing the highest with the lowest sex-specific ferritin quintile. The association remained significant after further controlling for BMI, hsCRP, adiponectin, and GGT. To update the evidence reported in previous meta-analyses, we searched all prospective studies evaluating the association between blood ferritin and incident diabetes on PubMed prior to October 24, 2012. Besides our prospective study, 9 additional studies were also included. The pooled RR was 1.60 (95% CI: 1.25, 2.04) when comparing the highest with the lowest category of ferritin with a moderate heterogeneity (I(2) = 49.0%; P = 0.03). A significant linear dose-response relationship was detected in this meta-analysis. Overall, our results indicate an independent and significant positive association between higher plasma ferritin, a marker of elevated body iron stores, and increased risk of developing type 2 diabetes in middle-aged and elderly Chinese adults, which is similar to Western populations.