Storm Xynthia (2010) was responsible for an important coastal flood on French Atlantic coast. This flood has impacted low-lying urbanized areas, which conducts to the death by drowning of 41 people who were trapped inside single-story houses without chances to escape. Review of the storm pointed out the recent growth of urbanization in potentially flooded area, non adaptation of constructions to the risk, and predominance of elderly as a combination of reasons to explain the human toll. Based on these facts and existing methods of mortality estimation (dutch example of mortality function), we propose an adapted method to predict potential number of death in case of coastal flood. To evaluate the potential number of death, this methodology uses information about: the location and configuration of residential houses (based on previous work, the VIE index) in order to asses constructions as a threat for people, the characteristics of population (age) on coastal municipalities in order to asses human vulnerability, and the occupation of houses depending if there are residential or secondary homes. Following the dutch mortality function, it allows to estimate a probability of death amongst a number of people exposed. The novelty is that we can propose a probability of death for each town depending of the above mentioned criteria. This work has been validated using information about storm Xynthia. It gives good correlation with the toll of this event (for three impacted towns of Baie de l’Aiguillon: La Faute-sur-Mer, l’Aiguillon-sur-Mer and Charron). It has been applied to a non impacted but potentially vulnerable area, Noirmoutier Island, to assess its vulnerability to futures coastal flood. The usefulness of the method is based on the possibility to use it in a Cost-Efficiency Analysis (CEA) in a way of supporting decision process at a local scale. It allows to compare different scenarios of vulnerability reduction for houses and to evaluate the number of death with and without measures. It gives a number of life potentially saved with the measure. The authors conducted this type of CEA comparing four vulnerability reduction measures. Presently applied to the two study areas, the CEA could be applied to other territories in order to identify the best solutions in order to reduce vulnerability for people, depending on local characteristics of population and danger of buildings. It will be useful for land use planning, for example for European flood directive.