Objective: Checking the sociodemographic and clinical factors associated with the prevalence of postpartum hemorrhage (PPH) in a maternity school. Materials and methods: A quantitative cross-sectional study in a maternity hospital in Salvador, Bahia, Brazil. In data collection, we used a standardized form that contained sociodemographic and clinical data from the medical records of 83 women for the period of 2018. Stata version 14 software was used in the analyses. Bivariate analysis was conducted using Pearson’s or Fisher’s exact tests. Poisson regression was performed with robust variation in multivariate analysis. Prevalence ratios (PR) and respective 95 % confidence intervals were estimated. The significance level of the tests was 5 %. Results: The prevalence of PPH was 38.6 % and 25.6 % for atony as the cause. In the bivariate analysis, there was an association between PPH and non-breastfeeding in the first hour of life (p = 0.039). In the multivariate analysis, it was identified that multiparous women had an increase in the prevalence of PPH by almost twice (PR = 1.97). Not breastfeeding in the first hour of life increased this prevalence more than four times (PR = 4.16). Conclusions: Monitoring multiparous women during birth care and encouraging breastfeeding in the first hour of life may decrease the prevalence of PPH.