Question Answering (QA) is a specialized form of information retrieval characterized by information needs that are expressed as natural language statements or questions. Query expansion is an approach which is used to extend question key words with new related words that are not included in question. In this study, anew association rule based question processing model is introduced. This model is used for expanding question keywords with the most related terms using new hybrid association rule base. This hybrid rule base takes into account both the exact match of association rules and the hierarchical match of semantic similarity to overcome the mismatch problem between questions and answer words. Therefore, it contains association rules between document words at a semantic level. The main goal of this new model is to enrich the question by reformulating it into similar meaning queries with additional information and hence for improving question answering process.