In ship operation, loss of propulsion is considered as a major hazard, which can put ship safety in jeopardy. However, due to the hostile and variable conditions under which marine engines operate, the existing barriers and maintenance strategies sometimes fall short in preventing breakdown or emergency repair scenarios. In this study, a Bayesian Network (BN) model has been designed to capture the relationships among various two-stroke engine's operational parameters in order to evaluate key performance characteristics influencing an engine health. The BN model is then fine-tuned to estimate the prior probabilities of its root nodes from engine operational parameters through the use of fuzzy logic. Operational inputs from engine operational data can be inserted into the fuzzy model and then output coupled with the BN aid to calculate the overall health of the engine. The advantage of this hybrid model is that it can effectively incorporate real time observable engine performance values to run simulations for trending various operational parameters for marine engine health diagnosis. The findings through a real cases reveals that the model enables swift performance assessment and helps ship operators frequently monitor the engine health.