This thesis addresses the challenge of route classification for LNG-powered ferries in the Baltic Sea, employing machine-learning clustering techniques to analyze historical AIS data. The study focuses on five LNG-powered ferries over a six-month period, aiming to classify their routes using clustering algorithms. The primary objective is to implement and evaluate two clustering algorithms, TimeSeriesKMeans with DTW and hierarchical-clustering with Frechet Distance, to determine their effectiveness in accurately identifying distinct routes. The secondary objective involves analyzing the operational impact of these routes. Our findings indicate that hierarchical-clustering with Frechet Distance consistently produced more distinct and meaningful clusters compared to TimeSeriesKMeans with DTW. However, hierarchical-clustering’s time complexity and computational demands are significantly higher. The analysis highlights the need for robust preprocessing techniques to handle AIS data anomalies and suggests that future work should consider larger datasets and alternative feature combinations to improve clustering outcomes and operational insights. This research contributes to the maritime industry’s efforts towards sustainability and operational efficiency by providing a methodology for route classification of the selected vessels, enhancing decision-making in route planning and operational analysis.