Keyword search has attracted a great deal of attention for retrieving XML data because it is a user-friendly mechanism. But Keyword search is hard to directly improve search quality because lots of keyword-matched nodes may not contribute to the results. XML includes rich semantic informationthese semantics are helpful to information retrieval process. This paper investigates the compelling problem of how to take advantage of XML semantics to improve keyword search quality. We design an XML keyword search approach, called XIOF, that can derive the keyword query and generate a set of effective structured queries by analyzing the given keyword query and the schemas of XML data sources. Furthermore, we provide an algorithm to computing the similarity score between two XIO. We have implemented our method on real datasets and the experimental results show that XIOF approach achieves both high recall and precise when compared with existing proposals.