Information and access to data on reproductive health is limited for Yemen, one of the least developed countries in the Middle East and with one of the highest birth rates in the world. This study uses primarily the 1997 Yemeni Demographic, Maternal and Child Health Survey (n=10,414 ever-married women age 15-49). We first examine levels, patterns, and differentials in maternal and infant health. Rates of infant and maternal mortality are very high (75/1000 and 351/100,000 live births, respectively). Most Yemeni women do not receive health care during pregnancy, 16% of deliveries occur in health facilities, and postpartum care is estimated among only 6% of deliveries. We then analyze patterns of maternal health care utilization, differentials in service coverage, and reasons for the underutilization of maternity services. Logistic regression analysis is used to assess the effects of environmental, socio-economic, and demographic variables for the probability of (1) receiving prenatal care; (2) receiving assistance during delivery; and (3) the place of delivery (home, public, or private health facility). Separate models are also estimated for urban and rural women. Among other known determinants of maternal health care utilization, qat chewing decreases the likelihood of receiving prenatal care, while less autonomous women are more likely to give birth at home but are not significantly less likely to receive prenatal care. The results are interpreted in light of additional information on reproductive health services in Yemen. Lastly, the implications of these findings for the potential improvement of services are considered.