BACKGROUND: Stillbirth rates in Afghanistan have declined little in the past decade with no data available on key risk factors. Health care utilisation and maternal complications are important factors influencing pregnancy outcomes but rarely captured for stillbirth in national surveys from low- and middle-income countries. The 2010 Afghanistan Mortality Survey (AMS) is one of few surveys with this information. METHODS: We used data from the 2010 AMS that included a full pregnancy history and verbal autopsy. Our sample included the most recent live birth or stillbirth of 13 834 women aged 12-49 years in the three years preceding the survey. Multivariable Poisson regression was used to identify sociodemographic, maternal, and health care utilisation risk factors for stillbirth. RESULTS: The risk of stillbirth was increased among women in the Central Highlands (aRR: 3.01, 95% CI 1.35, 6.70) and of Nuristani ethnicity (aRR: 9.15, 95% CI 2.95, 28.74). Women who did not receive antenatal care had three times increased risk of stillbirth (aRR: 3.03, 95% CI 1.73, 5.30), while high-quality antenatal care was important for reducing the risk of intrapartum stillbirth. Bleeding, infection, headache, and reduced fetal movements were antenatal complications strongly associated with stillbirth. Reduced fetal movements in the delivery period increased stillbirth risk by almost seven (aRR: 6.82, 95% CI 4.20, 11.10). Facility births had a higher risk of stillbirths overall (aRR: 1.55, 95% CI 1.12, 2.16), but not for intrapartum stillbirths. CONCLUSIONS: Targeted interventions are needed to improve access and utilisation of services for high-risk groups. Early detection of complications through improved quality of antenatal and obstetric care is imperative. We demonstrate the potential of household surveys to provide country-specific evidence on stillbirth risk factors for LMICs where data are lacking.