This research work tries to provide a model-based, distributed Fault Diagnosis (FD) framework that will eventually act as an early warning system for online monitoring of the marine propulsion process, in order to avoid future failures and accidents. Furthermore, the introduction of adaptive thresholds in the monitoring modules of the monitoring agents helps minimize false alarms and reduces the conservativeness in decision-making. Finally, this work comes to provide an integrated sensor and process FD framework that can be adopted on-board marine vessels in the future, and narrow down the gap, regarding FD for marine vessels which is closely linked with maritime safety and maintenance decision-making.