<p>Cargo-related fires on container vessels have become a critical safety concern in maritime transportation due to their increasing frequency and the severe human, environmental, and economic losses they cause. The multidimensional nature of these incidents requires that fire risk be addressed not only in terms of technical failures but also in relation to system-level weaknesses in human, organisational, and operational processes. The aim of this study is to classify the causes of cargo-related fires in the cargo areas of container vessels using the Human Factors Analysis and Classification System (HFACS) and to identify critical risk patterns using Association Rule Mining (ARM) techniques. Full-text accident investigation reports obtained from official databases were systematically coded according to the HFACS-PV framework. The resulting qualitative data were transformed into a binary dataset and analysed using the Apriori and Predictive Apriori algorithms. The results indicate that misdeclaration and misclassification of cargo under the IMDG Code, inadequate documentation control, stowage planning errors, and deficiencies in information sharing are key contributors to fire initiation. Furthermore, reduced situational awareness, communication failures, and delayed response behaviours were found to intensify fire severity under adverse operational conditions.</p>