Marine and Offshore renewable structures are widely being developed across the globe, and specifically in Australia, under controlled conditions with little focus on having advanced health monitoring systems, optimization of the structures design and operation to undertake complex operations. Considering a data-driven approach to health management is an essential part of the future offshore renewable energy industries. It allows the detection of critical faults and assists in estimating the remaining useful lifetime (RUL) facilitating more reliable offshore renewable structures. Despite the advantage of machine learning techniques in asset health condition monitoring, few studies have considered the integration of such prognostics into maintenance planning for remotely operated offshore facilities. This paper proposes a model to help the marine and offshore industry in real-time maintenance planning. The model will consider the imperfect RUL prognostics incorporating higher degrees of uncertainty involved with remotely operating systems in offshore environments. A Bayesian Meta-Model is developed for time series change point (CP) detection of multiple anomalies in a random process. The proposed methodology will be useful in incipient failure detection and maintenance planning of complex systems operating in harsh and highly random marine environments.