The aim of this study was to assess temporal trends in incidence and underlying causes of maternal deaths from obstetric hemorrhage in France and to describe clinical care before and after implementation of the first national guidelines published in 2004 and updated in 2014. Data from all hemorrhage-related maternal deaths between 2001 and 2015 were extracted from the French Confidential Enquiry into Maternal Deaths. We compared the maternal mortality ratio (MMR), cause of obstetric hemorrhage, and death preventability by triennium. Critical care, transfusion, and obstetric management among women who died were described for 2001 to 2003 and 2013 to 2015. The MMR from obstetric hemorrhage significantly decreased over time from 2.3 of 100,000 livebirths (54 of 2,391,551) in 2001 to 2003 to 0.8 of 100,000 livebirths (19 of 2,412,720) in 2013 to 2015. In 2001 to 2003, uterine atony accounted for 50% (27 of 54) of maternal deaths vs 21% (4 of 19) in 2013 to 2015. As compared to 2001 to 2003, an increased proportion of women had hemodynamic continuous monitoring in 2013 to 2015 (30%, 9 of 30, vs 47%, 8 of 18) and received vasopressor infusion therapy (57%, 17 of 30, vs 72%, 13 of 18), and a smaller proportion was extubated during active hemorrhage (17%, 5 of 30, vs 0 of 18). Transfusion therapy was initiated more frequently and earlier in 2013 to 2015 (71 vs 58 minutes). In 2013 to 2015, 88% of maternal deaths due to hemorrhage remained preventable. The main identified improvable care factors were related to delays in diagnosis and surgical management, particularly after cesarean delivery. Maternal mortality by obstetric hemorrhage decreased dramatically in France between 2001 and 2015, particularly mortality due to uterine atony. Among women who died, we detected fewer instances of substandard transfusion management or critical care. Nevertheless, opportunities for improvement were observed in most of the recent cases.