An ontology-based approach to effectively process approximate query for multi-source and heterogeneous XML data is proposed

in which the traditional approximate XML query strategy using tree pattern relaxation is integrated with the idea of ontology-based XML data integration. Firstly

a mapping model(PTO)from document type definition(DTD)structure to global schema is constructed

then PTO is combined with relaxation and scoring mechanism and a novel algorithm-OAXQ is presented. Thus

user query language based on global schema becomes a simple subset of object query language instead of Xpath expressions

and relaxations are accomplished by a series of conversion of mapping rules instead of transform of query tree patterns. Experiments of intelligent design of automobile shape show that comparing with SSO algorithm the query speed and precision is increased by 20 times and 2%-24% respectively by OAXQ