To investigate the correlation between first trimester vaginal bleeding and preterm birth (PB), and to offer suggestions on the perinatal health care and preterm birth prevention. A birth cohort study was conducted on 10 179 pregnant women. Unconditional logistic regression model was used to evaluate the associations between vaginal bleeding and preterm birth in sub-preterm groups. Of the 10 179 pregnant women included, a total of 1001 women suffered from vaginal bleeding during the first trimester, of which 119 suffered from PB. Any vaginal bleeding increased the risk of PB. Severe bleeding was a high-risk factor of PB, associated with 4.8-fold risk of very PB, 2.7-fold risk of spontaneous PB without PROM (premature rupture of membrane) and 4.6-fold risk of medical induced PB. Bleeding prolonged more than 1 week increased 66% risk of PB and 36% risk of PB on initial episode happened in 5-12 weeks of gestation age, especially in moderate PB, in medical-induced PB and in spontaneous PB with PPROM (preterm premature rupture of membrane which is one cause of PB). Mild bleeding or bleeding within 1 week or initial episode happened within 4 weeks of gestation age possibly had no influence on PB. Vaginal bleeding in the first-trimester was an independent risk factor for PB. The severity, duration and initial time of vaginal bleeding had different effects on different subtypes of PB.