There has been increased interest in and experimentation with demand-side mechanisms such as the use of vouchers that place purchasing power in the hands of targeted consumers to improve the uptake of healthcare services in low-income settings. A key measure of the success of such interventions is the extent to which the programmes have succeeded in reaching the target populations. This article estimates the coverage of facility deliveries by a maternal health voucher programme in South-western Uganda and examines whether such coverage is correlated with district-level characteristics such as poverty density and the number of contracted facilities. Analysis entails estimating the voucher coverage of health facility deliveries among the general population and poor population (PP) using programme data for 2010, which was the most complete calendar year of implementation of the Uganda safe motherhood (SM) voucher programme. The results show that: (1) the programme paid for 38% of estimated deliveries among the PP in the targeted districts, (2) there was a significant negative correlation between the poverty density in a district and proportions of births to poor women that were covered by the programme and (3) improving coverage of health facility deliveries for poor women is dependent upon increasing the sales and redemption rates. The findings suggest that to the extent that the programme stimulated demand for SM services by new users, it has the potential of increasing facility-based births among poor women in the region. In addition, the significant negative correlation between the poverty density and the proportions of facility-based births to poor women that are covered by the voucher programme suggests that there is need to increase both voucher sales and the rate of redemption to improve coverage in districts with high levels of poverty.