Some issues of increasing the energy efficiency of ships by improving navigation methodsThis paper presents a comprehensive analysis of contemporary trends in automatic ship control systems with particular emphasis on artificial intelligence technology integration in autopilots, propulsion systems, and energy efficiency enhancement in maritime transport. The study covers the evolution from classical PID controllers to intelligent control systems, including neural networks of various architectures, fuzzy logic, adaptive algorithms, and modern machine learning systems. Special attention is given to systematic analysis of AI technology applications for ship course keeping tasks, automatic trajectory control in complex navigation conditions, propulsion system optimization, and comprehensive optimization of ship system energy consumption. The advantages and fundamental limitations of various approaches to intelligent ship control are thoroughly examined, along with their impact on maritime safety, economic efficiency of maritime transport, and environmental aspects of shipping. A deep analysis of autonomous navigation development prospects and the critical role of AI in creating intelligent maritime transport systems of the future is conducted. The research includes comparative analysis of traditional propeller installations and azimuthal propulsion complexes, modern developments in energy-saving devices such as Becker Mewis Ducts, and integration of adaptive control systems with propulsion technologies. Results demonstrate that integration of advanced AI technologies in autopilot systems enables achieving significant improvements in course control accuracy by 25-35%, substantial fuel consumption reduction by 10-15%, and qualitative enhancement of overall maritime safety. Azimuthal thrusters provide 32% reduction in incident rates and 33-67% improvement in maneuvering characteristics compared to traditional systems. Energy-saving devices achieve fuel savings up to 8% for slow full-form vessels. The work systematizes and critically analyzes results of modern research on fuzzy controllers, neural network autopilots of various architectures, hybrid ANFIS systems, backstepping control methods, LSTM networks for trajectory prediction, reinforcement learning, event-triggered approaches, and predictive control technologies.