Abstract: Objectives In the field of ship design, cabin spatial layout is not only crucial for enhancing navigation safety and operational efficiency, but also a key factor in crew comfort and overall vessel performance. This paper aims to establish a multidimensional evaluation system for cabin layout optimization, systematically review relevant optimization algorithms, analyze current technical bottlenecks, and propose future research directions to provide theoretical and technical support for intelligent cabin design in ships. Methods First, a multidimensional evaluation system tailored for ship cabin spatial layout optimization was proposed, comprehensively considering factors such as functional relevance, human activities, and environmental aspects, thereby providing clear objectives and direction for the selection and design of subsequent optimization algorithms. Then, this paper systematically reviews important algorithms proposed by researchers in recent years for ship cabin layout optimization, including but not limited to genetic algorithms, gravitational search algorithms, integer programming methods, and group motion simulation-driven approaches, summarizing research achievements in obtaining high-quality spatial layouts using these optimization methods. Furthermore, a support relationship table between optimization methods and evaluation metrics was constructed, and mainstream methods were compared and analyzed from multiple dimensions. Simultaneously, for complex multi-deck scenarios, differences between single- and multi-deck layouts were compared, representative examples were summarized, and existing shortcomings were analyzed. Results Despite significant progress made by various algorithms in the domain of ship cabin spatial layout optimization, numerous limitations persist, such as generally neglecting problem considerations at the micro-execution level and difficulties in efficiently modeling multi-deck layouts. Conclusions Finally, this paper proposes future research directions from multiple perspectives, including the construction of domain-specific datasets, the development of innovative intelligent optimization algorithms, and the design of human-in-the-loop deep reinforcement learning frameworks.