The accuracy of perinatal and infant mortality rates in most developing countries is questionable. We measured perinatal and infant mortality rates in a rural district of Thailand and compared them with the official statistics to assess accuracy. All births and infant deaths in a rural district of Thailand over a one-year period were surveyed. The corresponding official statistics were also collected. The mothers or the relatives of all stillbirths and infant deaths were interviewed about the registration of the stillbirths or infant deaths. The surveyed perinatal and infant mortality rates were 22.0 and 23.1 respectively. The under-registration of stillbirths was 100% and for infant deaths 45%. All the non registered infant deaths were in situations in which the infant died before the registration of birth. These results document the degree and nature of under-reporting of perinatal and infant mortality in rural Thailand.