To investigate uptake and provision of antenatal care (ANC) services in the Uzazi Bora project: a demonstration-intervention project for Safe Motherhood and prevention of mother-to-child transmission of HIV in Kenya. Data were extracted from antenatal clinic, laboratory and maternity ward registers of all pregnant women attending ANC from January 2004 until September 2006 at three antenatal clinics in Mombasa and two in rural Kwale district of Coast Province, Kenya (n = 25 364). Multiple logistic and proportional odds logistic regression analyses assessed changes over time, and determinants of the frequency and timing of ANC visits, uptake of HIV testing, and provision of iron sulphate, folate and single-dose nevirapine (sd-NVP). About half of women in rural and urban settings (52.2% and 49.2%, respectively) attended antenatal clinics only once. Lower parity, urban setting, older age and having received iron sulphate and folate supplements during the first ANC visit were independent predictors of more frequent visits. The first ANC visit occurred after 28 weeks of pregnancy for 30% (5894/19 432) of women. By mid-2006, provision of nevirapine to HIV-positive women had increased from 32.5% and 11.7% in rural and urban clinics, to 67.0% and 74.6%, respectively. Equally marked improvements were observed in the uptake of HIV testing and the provision of iron sulphate and folate. Provision of ANC services, including sd-NVP, increased markedly over time. While further improvements in quality are necessary, particular attention is needed to implement evidence-based interventions to alter ANC utilization patterns. Encouragingly, improved provision of basic essential obstetric care may increase attendance.