A method and system for developing fault models from unstructured text documents, such as text verbatim descriptions from customers and service technicians. An ontology, or data model, and heuristic rules are used to identify and extract descriptive terms from the text verbatim document. The descriptive terms are then classified into types, including symptoms, failure modes, and parts. Like-meaning but differently-worded descriptive terms are then merged using text similarity scoring techniques. The resultant symptoms, failure modes, parts, and correlations are then assembled into a fault model, which can be used for real-time fault diagnosis onboard a vehicle, or off-board at service shops.