The research supports the thesis that a significant improvement in the management of uncertainty characterising the assessment of climate change policies can come from the implementation of a Bayesian network (BN) approach, that allows the user to characterise, incorporate and communicate the uncertainty. The main goal of an analysis of climate change and uncertainty should look at the formulation of optimal climate policies taking into account how uncertainties, concerning not only environmental systems and dynamics but also socio-economic factors, such as technological change and human adjustments to climate change impacts, might affect decisions about policy strategies. BNs are a new generation of systems that are capable of modelling real-world decision problem using theoretically sound methods of probability theory and decision theory. Based on graphical representation of the problem structure, these systems allow for combining expert opinions with frequency data. In order to investigate its potentials and limits, a BN model was implemented and tested on a specific case study, to analyse different adaptation options to the impacts of sea level rise. Uncertainties were incorporated and reduced through the use of expert subjective probabilistic judgments and Bayesian learning. The BN provided a support for adaptive management, structuring an informed decision making process, with a precautionary approach that overcame the uncertainty of future projections and models’ estimates.