Climate change is expected to affect health through changes in exposure to weather disasters. Vulnerability to coastal flooding has decreased in recent decades but remains disproportionately high in low-income countries. We developed a new statistical model for estimating future storm surge-attributable mortality. The model accounts for sea-level rise and socioeconomic change, and allows for an initial increase in risk as low-income countries develop. We used observed disaster mortality data to fit the model, splitting the dataset to allow the use of a longer time-series of high intensity, high mortality but infrequent events. The model could not be validated due to a lack of data. However, model fit suggests it may make reasonable estimates of log mortality risk but that mortality estimates are unreliable. We made future projections with and without climate change (A1B) and sea-based adaptation, but given the lack of model validation we interpret the results qualitatively. In low-income countries, risk initially increases with development up to mid-century before decreasing. If implemented, sea-based adaptation reduces climate-associated mortality in some regions, but in others mortality remains high. These patterns reinforce the importance of implementing disaster risk reduction strategies now. Further, while average mortality changes discontinuously over time, vulnerability and risk are evolving conditions of everyday life shaped by socioeconomic processes. Given this, and the apparent importance of socioeconomic factors that condition risk in our projections, we suggest future models should focus on estimating risk rather than mortality. This would strengthen the knowledge base for averting future storm surge-attributable health impacts.