Future sea-level rise will increase coastal flood risk in the U.K., yet the hazard uncertainties associated with such future risk estimates have not been fully explored. The sensitivity of coastal flood-risk mapping to future uncertainties was investigated by propagating ranges of plausible parameters through a LISFLOOD inundation model of a significant historic flood event to the North Somerset (U.K.) coast. Mean sea-level rise (including land movement) was found to have the greatest effect on the extent of flood inundation. Analysis of the latest research into the future storm-surge climate of the U.K. indicates no change above natural variability, thus, future, extreme water-level estimates (for the U.K.) should be based on observations and not Regional Circulation Models until research indicates otherwise. Evidence suggests that the current approach of forcing the inundation model with an extreme water level of a constant return period is incorrect. This uncertainty of the peak storm tide height along the coastline had a significant effect on our results. We present a new boundary-forcing technique to force the inundation model with (method C), based on the spatial characteristics of real events, which can account for the natural storm-surge variability. Indeed, if sea-level rise is included with method C, a great deal of the uncertainty surrounding such a future flood-hazard estimate can be quantified and communicated clearly and effectively.