To address the issues of information silos, traceability difficulties, and decision-making delays in the construction management of expressway projects, this study proposes a novel AI-driven traceability management framework based on the deep integration of BIM and WBS. Innovatively, a bidirectional semantic fusion model is introduced to dynamically map component IDs with work package codes, alongside an intelligent early warning mechanism. The framework integrates multi-source heterogeneous data and big data analytics to support a closed-loop process spanning "design-construction-operation." Using the Nanping section of the Shanan Expressway in Fujian Province as a case study, empirical applications were conducted across four dimensions: schedule, quality, cost, and management expenses, with comparisons made against traditional management approaches. The digitalization rate of project completion documents has reached 85%, reducing the manual input workload by 80% and significantly enhancing the efficiency and accuracy of construction management.