This paper presents the design, implementation, and comprehensive evaluation of an intelligent virtual simulation platform deeply integrated with a Case-Based Learning (CBL) pedagogy, specifically developed for automotive electrical fault diagnosis training. The research addresses three critical challenges in traditional vocational training: high equipment costs, safety risks associated with physical fault insertion, and the limited diversity of fault scenarios available for practice. We developed a novel, data-driven virtual simulation system featuring three core modules: 1) a high-fidelity 3D interactive model library enabling immersive disassembly and assembly of vehicle components, 2) an AI-assisted intelligent fault case generation and management system (IFCGMS) that supports configurable compound faults through a rule-based reasoning and fault tree analysis model, and 3) a dynamic principle visualization and intelligent assessment module (DPVIAM) that tracks, evaluates, and provides real-time feedback on student operations. A quasi-experimental study was conducted, comparing an experimental group (N=33) using the integrated “virtual simulation-CBL” model against a control group (N=30) receiving traditional instruction. The results demonstrate a significant superiority of the experimental group, with a +12.2 point improvement in real-vehicle troubleshooting scores and a +9.6 point gain in comprehensive course performance. Furthermore, in-depth analysis of platform log data using formal computational models—including a Diagnostic Sequence Similarity Score (DSSS) to evaluate diagnostic procedures and Pearson correlation analysis (r = 0.82) to link virtual and real performance—revealed that diagnostic efficiency and error correction patterns in the virtual environment are strong predictors of final practical competency. This work contributes a scalable, computer-aided instructional framework that effectively leverages virtual simulation, pedagogical design, and learning analytics, offering a replicable solution for enhancing skill acquisition in high-risk, high-cost technical training domains such as automotive diagnostics.