An integrated approach combining Bayesian Network with GIS was developed for making a probabilistic prediction of sea level rise induced coastal erosion and assessing the implications of adaptation measures. The Bayesian Network integrates extensive qualitative and quantitative information into a single probabilistic model while GIS explicitly deals with spatial data for inputting, storing, analysing and mapping. The integration of the Bayesian Network with GIS using a cell-by-cell comparison technique (aka map algebra) provides a new tool to perform the probabilistic spatial analysis. The spatial Bayesian Network was utilised for predicting coastal erosion scenarios at the case study location of Tanna Island, Vanuatu in the South Pacific. Based on the Bayesian Network model, a rate of the island shoreline change was predicted probabilistically for each shoreline segment, which was transferred into GIS for visualisation purposes. The spatial distribution of shoreline change prediction results for various sea level rise scenarios was mapped. The outcomes of this work support risk-based adaptation planning and will be further developed to enable the incorporation of high resolution coastal process models, thereby supporting localised land use planning decisions.