To deal with the problem of approximate query against XML documents, based on the extensions of document attribute units, the authors propose a novel XML approximate query approach which can provide the relevant query results to the user’s original query. The leaf nodes and attribute nodes of XML documents are treated as attribute units. Then based on the concept of the agree set, the maximum set is exported and the minimum nontrivial functional dependence sets are generated consequently. Thus the approximate dependence relations can be found. By using the approximate dependence relations, the approximate candidate keys and approximate keywords are found. After that, this approach ranks the attribute units according to their supported degree and expands the original query by regarding the importance sequence of attribute units. The first attribute unit to be relaxed must be the least important attribute unit and has the maximum relaxation degree. The relaxed query is used to query the XML documents and the relevant query results of the original query are obtained. The experimental results and analysis demonstrate that the XML approximate query approach presented can efficiently meet the user’s query intentions and has a high performance as well.