In this paper a statistical method is described for predicting fatigue accumulation. The proposed method for predicting accumulated fatigue is based on a nonparametric Bayesian analysis. Bayesian analysis provides a natural prediction framework for combining long-term design information and actual measured data. This work is part of the Monitas II project, which delivered an advisory hull monitoring system for FPSOs with automatic data analyses capabilities. The prediction tool is to be embedded in this hull monitoring system.