Recovering traceability is an important software goal in the reverse engineering process. Conventional methods based on the information retrieval model have recovered traceability links for supporting software engineers. In such methods, the call relations in the source code directly impact the accuracy of traceability links in large projects, e.g. enterprise projects. This paper uses a source code parser to create a syntax tree as a preprocessing step for information retrieval to recover traceability links between source code and documents. The effectiveness of our proposal is confirmed by a comparison with real product data obtained from an enterprise project.