As countries continue making gains towards the attainment of sustainable development goals, surveillance of health outcomes at the sub-district level will become important as district level indicators may mask areas where progress is slow. To achieve this high level of surveillance, it may be necessary to pool data from multiple data sources with different spatial resolutions. The aim of this study was to estimate and model under-five mortality risk at the sub-district level in Malawi by combining multiple data sources. We used Bayesian hierarchical models to combine the Demographic and Health Survey (DHS) data with Census data in a principled framework. A binomial generalized linear geostatistical model was fitted to estimate the risk of under-five mortality in the presence of the various covariates. Results showed that mother’s age and weight of child at birth were associated with under-five mortality. However, the posterior odds showed no significant differences in dying for children from mothers across different ages. In addition, the results showed that the risk of under-five mortality is higher in the northern region and along lakeshores as well as districts in the lower Shire. The study provided a means for performing small area estimation of population parameters of interest. In addition, using survey findings along with risk maps is essential for disease monitoring and surveillance purposes as well as for strengthening survey findings. More importantly, the project has improved our understanding of methods used in combining information from different sources.