Background Parity, pregnancy loss, and breastfeeding duration were found to be associated with diabetes. However, the results are inconsistent. Also, no epidemiological studies have examined the association of these reproductive factors with diabetes in the same large population. We aim to investigate the associations between parity, pregnancy loss, breastfeeding duration, and the risk of maternal diabetes in middle-aged and elderly Chinese females. Methods We included 131 174 females aged >= 40 years from the REACTION study (Risk Evaluation of Cancers in Chinese Diabetic Individuals: A Longitudinal Study). Multivariable linear regression and logistic regression were used to assess the association between parity, pregnancy loss, and breastfeeding duration and type 2 diabetes. Results The number of parities and breastfeeding duration were positively related to fasting plasma glucose, 2-hour postload glucose, glycosylated hemoglobin, and homeostatic model assessment of insulin resistance. Compared with those with one birth, nulliparous women or women with 2 or >= 3 births had a significantly increased risk of diabetes. The odds ratios (OR) and 95% confidence intervals (CI) were 1.27 (1.10-1.48), 1.17 (1.12-1.22), and 1.28 (1.21-1.35), respectively. Compared with women without pregnancy loss, those who underwent 2 (OR 1.09; 95% CI, 1.04-1.14) or >= 3 pregnancy losses (OR 1.11; 95% CI, 1.04-1.18) had an increased risk of diabetes. Moreover, women with a breastfeeding duration >= 0 to 6 months (OR 0.82; 95% CI, 0.75-0.90) and >= 6 to 12 months (OR 0.94; 95% CI, 0.89-0.99) had a significantly lower risk of diabetes. Conclusions Nulliparous women or women with multiparity or more than one pregnancy loss have an increased risk of diabetes in later life, while women who breastfeed more than 0 to 12 months have a lower risk of diabetes.