Intelligent ships, as key drivers of the shipping industry’s intelligent transformation, improve navigation safety and operational efficiency. It simultaneously introduces novel risks arising from the interplay of environmental, managerial, hardware, and intelligent system factors. This paper systematically reviews research progress in the identification and assessment of navigation risks for intelligent ships. First, it categorizes typical risk factors across four dimensions---environment, management, hardware, and navigation systems---highlighting their complex characteristics of “traditional–new” interdependence and “hardware–software” synergy. Second, the principles, applicability, and limitations of qualitative, quantitative, and comprehensive evaluation methods are critically examined, covering approaches such as Effects Analysis (FMEA), Event Tree/Fault Tree Analysis (ETA/FTA), Analytic Hierarchy Process–Fuzzy Comprehensive Evaluation (AHP-FCE), Bayesian Networks (BNs), and Evidence Theory. Furthermore, key challenges in data quality, regulatory adaptation, and human–machine collaboration are analyzed. Finally, future directions are outlined, including digital twin-driven real-time risk monitoring, generative artificial intelligence-enabled self-learning assessment, multi-ship collaborative decision-making, and standardized evaluation systems. This review aims to provide theoretical support for establishing a scientific, dynamic, and interpretable risk assessment framework for intelligent ship navigation.