The explicit safety knowledge embedded in ship navigation rules in textual form is crucial for maritime traffic safety management. With the advancement of intelligent shipping, rule understanding has rapidly developed to support smart navigation. However, the structural specificity and semantic contextualization of rule texts present challenges for machine interpretation. This study proposes an autonomous interpretation method for navigation rules under a “situation constraint – operational rules” framework. This method integrates Named Entity Recognition (NER) and Semantic Role Labeling (SRL) information extraction models to extract knowledge from ship navigation rules and represent it within a Neo4j-based knowledge graph. In an experimental evaluation using Tianjin Port navigation rules, the method successfully identified 125 rules from the test data, demonstrating its effectiveness.