Objectives: Intergenerational "mismatch" and/or growth conditions may be relevant to the epidemic of diabetes in developing populations. In a rapidly developing southern Chinese population, we tested whether maternal environment, proxied by maternal literacy, or family socio-economic position (SEP), proxied by paternal literacy, were associated with fasting glucose and diabetes. To assess if intergenerational mismatch contributed, we tested whether the associations varied by life course SEP. Methods: In 19,818 older (>= 50 years) adults from the Guangzhou Biobank Cohort Study (phases 2 and 3), we used censored and logistic regression to assess the associations of maternal and paternal literacy with fasting glucose, elevated fasting glucose and diabetes and whether these associations varied by sex, age or life course SEP. Results: Maternal, but not paternal, literacy was negatively associated with fasting plasma glucose (beta-coefficient -0.06 mmol/l, 95% confidence interval (CI) -0.11 to -0.01) and elevated fasting glucose (odds ratio (OR) 0.92, 95% CI 0.86-0.99) adjusted for age, sex, study phase, life course SEP, childhood growth, adiposity, number of offspring, and birth order. Associations of maternal and paternal literacy with fasting glucose, elevated fasting glucose and diabetes did not vary by sex, age or life course SEP. Conclusion: Offspring of literate mothers had lower risk for impaired glucose tolerance than offspring of illiterate mothers. Being raised by literate mothers may increase the likelihood of children with higher SEP and lower long-term disease risk, or better maternal conditions over generations may be associated with lower fasting glucose. Am. J. Hum. Biol. 23: 747-754, 2011. (C) 2011 Wiley Periodicals, Inc.