The Zapatista armed conflict began in the state of Chiapas, Mexico, in 1994, and overlaps pre-existing local disputes about land, religion, and other issues. Related disruptions in access to and utilization of health services have been alleged to have compromised local health status, particularly in vulnerable subgroups such as indigenous women and infants. The study objective was to measure maternal and perinatal mortality ratios and utilization of pregnancy-related health services in the region affected by the Zapatista conflict, and to describe associations between these primary outcome measures, socioeconomic and demographic factors, and factors associated with inter-party and intra-community conflict. A cross-sectional, population-based survey was conducted in 46 communities in three regions. The study subjects were 1227 women, 13-49 years old, who had been pregnant during the preceding 2 years (1999-2001). Principal outcome measures were maternal and perinatal mortality, and site of delivery. Secondary analyses explored associations between primary outcomes and socioeconomic, demographic, and conflict-related factors. Most births (87.1%) occurred at home. The crude observed maternal and perinatal mortality ratios were 607/100,000 and 23.5/1000 live births, respectively. Those who died had difficulty accessing emergency obstetrical care. Both home birth and mortality were associated with descriptors of intra-community conflict. Observed maternal and perinatal mortality ratios were substantially higher than those officially reported for Mexico or Chiapas. Reduction of high reproductive mortality ratios will require attention to socioeconomic and conflict-related problems, in addition to improved access to emergency obstetrical services.