Coastal flood model predictions are undeniably important for informing coastal management, but contain considerable uncertainty related to model structure, parameterisation, and input data. With these predictions becoming increasingly available through online flood maps, the uncertainty in these predictions presents considerable risks related to property devaluation. Such risks relate to real estate demand, measured by location preferences and willingness-to-pay (WTP) to buy and rent properties, based on access to flood predictions. Here, we evaluate the impact of flood predictions on coastal real estate demand in the UK by adopting an interdisciplinary approach, involving coastal flood modelling, a novel experimental WTP real estate survey of UK residents in response to flood model outputs, statistical modelling, and geospatial analysis. Our findings show that access to flood predictions dominates coastal real estate demand decisions in relation to personal preferences for location aesthetics, reflecting a shift in demand towards risk-averse locations. We also find that people do not consider flood prediction uncertainty in their real estate decisions, possibly due to an inability to perceive such uncertainty. Therefore, we argue that caution is needed when communicating flood predictions and using these predictions to inform coastal management. We advocate for the need to get flood models ‘right’ but recognise that this is a contentious issue as it implies having an error-free model, which is near impossible. Hence, we place greater emphasis on effectively communicating coastal flood predictions and their uncertainty to minimise real estate risks.