Evidence that childbearing is associated with future development of diabetes remains conflicting and the role of pregnancy loss in this association has not been investigated. We aimed to examine whether pregnancy and/or pregnancy loss (miscarriage, abortion, or stillbirth) are associated with maternal higher risk of diabetes later in life, using a population-based prospective cohort study (mean follow-up = 10.7 years), including 13,612 women (aged 35-65 at baseline). We found pregnancy per se did not change the risk of diabetes after considering the effect of education, smoking, alcohol consumption, physical activity, BMI, waist/hip ratio, hypertension, and hyperlipidemia (fully-adjusted OR: 1.04, 95 % CI: 0.82-1.31). Having more than four live births was associated with around two times higher risk of diabetes later in life (fully-adjusted OR: 1.77, 95 % CI: 1.12-2.80). Having more than two miscarriages was associated with about two-fold higher risk of diabetes (fully-adjusted Odd ratio (OR): 1.85, 95 % CI: 1.17-2.93). After further adjustment for parity, the higher risk of diabetes in those who had history of more than two miscarriages did not change substantially (OR: 1.82; 95 % CI: 1.15-2.88), but the association between more than four live births and diabetes disappeared when the role of pregnancy loss was considered (fully-adjusted HR: 1.06; 95 % CI: 0.54-2.08). No significant association was found between abortion, stillbirth and risk of maternal diabetes. Pregnancy per se did not increase risk of diabetes. Women who experience more than two miscarriages are at around two times higher risk of diabetes later in life. The association between high parity and diabetes is mediated by history of miscarriages and known risk factors of diabetes. The underlying reason for association between miscarriage and diabetes needs further investigation.