CoPropel is a Horizon Europe funded project which aims to develop marine propellers for cargo vessels that are more fuel and cost efficient to operate than their more common metallic counterparts. One of the areas the project aims to tackle is the maintenance aspect of marine propellers, which can be very cost intensive. This is due to the difficulty of inspection or maintenance since the components are underwater and quite significant in size, which may lead to the need for divers or dry-dock time. The way the project aims to address this is by taking advantage of the non-monolithic construction of the composite propeller to integrate a Structural Health Monitoring (SHM) system to monitor the strain, and subsequently infer the condition, of the propeller during operation without the need to stop for inspection. This paper explores the challenges encountered in the development of the SHM system, including: 1.) inspection requirements as set out by existing guidelines and regulations, 2.) sensor integration in the composite structure, 3.) data transmission from a rotating underwater component to a data acquisition system within the ship and 4.) usage of the acquired data for maintenance decision making. Feasible alternative systems are explored including Rayleigh Backscatter based Fibre Optic Systems (FOS) for distributed strain sensing as well as a strain gauge system with wireless underwater data transmission. Preliminary small scale underwater tests are conducted to evaluate the performance of the explored concepts under simulated operational conditions. Finally, we combine the strain data acquired from the SHM system with numerical models of the propeller through a multifidelity approach. The first aim is to create a better spatial distribution of the strain measurements across the propeller by fusing data from sensors covering a limited area, with strain data from numerical models that have better distribution across the propeller. The second goal is to provide digital twinning capability through the use of the numerical model to provide extrapolated estimates (calibrated by live sensor data) of maintenance requirements (i.e. remaining life) based on current and hypothetical operational profiles to better inform operators when making future operational decisions. Through the combination of the SHM system and digital twinning capabilities we provide increased data driven decision making capabilities to better improve operational efficiency and maintenance costs.