Recently, with the emergence of climate change and environmental pollution issues due to global warming, the importance of renewable energy generation has been highlighted. As demand increases, many of these systems are being installed offshore, where space utilization is more feasible. However, various environmental loads such as waves, wind, and currents act on these systems, and access is more difficult compared to onshore installations, resulting in high maintenance costs. Recently, to overcome these challenges, research has been conducted on implementing digital twins. This involves synchronizing sensor data measured from actual operating structures with virtual models in real-time, allowing users to monitor desired responses and make quick maintenance decisions. To prevent accidents, it is crucial for virtual models to compute responses faster using real-time measurement data. In this study, fluid-structure coupled analysis was performed on multi-linked floating offshore structures. And then, transfer functions between virtual sensors and structural responses at locations where significant stresses occur were calculated. These transfer functions were derived to predict structural responses using measured sensor data obtained from model tests, and we validated the application of this prediction method. Furthermore, it was evaluated the performance of the prediction method by comparing its results with measured data. Ultimately, we confirmed predicting structural responses with less than about 3% error compared to measured data, at computational speeds of less than 1 s.