Introduction. Maternal mortality is an indicator of the quality of obstetric care. In 1996, France established a surveillance system that includes monitoring rates and causes, as well as a confidential survey to investigate them. Methods. The definitions used are those of the WHO. An initial analysis focuses on data derived from vital statistics, particularly rates (ratio of obstetric deaths to live births); the second analysis is based on the Confidential Survey of Maternal Deaths and their review by the National Committee of Experts (percentage of preventability). Results. The rate is estimated at between 9 and 13 deaths per 100,000 live births. Following a decline recorded between 1996 and 2000, the most recent trend is less favorable. The rate increases with age (the risk is 8 times higher at age 40 than at ages 20–24) and by nationality (the rate is twice as high among non-European women). France ranks in the middle among comparable countries and could do better, following the example of Finland (6 per 100,000). Eighty-four percent of deaths occur in a public hospital, 10% in a private clinic, and 6% at home. The leading cause is hemorrhage (21% of maternal deaths), primarily postpartum, followed by complications of hypertension (12%), then amniotic fluid embolism (7%). More than half of maternal deaths from direct obstetric causes are considered preventable (73% of hemorrhages, 71% of infections, but only 43% of HTN). Conclusion. There is still significant room for improvement, both in identifying the causes of this tragedy and addressing them, and in optimizing care, since European countries comparable to ours continue to perform better. (R.A.)