In this paper, an ontologybased Internet search model is designed. This model is proposed in the background of the Internet's mass, changeable resources.On the one hand, its semantic inference solves the problems, such as the low recall, which appears in the search; on the other hand, it makes good use of ontology resources to do query excitation that provides more userfriendly search and advances the efficiency of the ontological inference. This model is validated in the case of the stock domain ontology, and the results prove that the model has more flexible and effective semantic search characteristics for the complexstructured and typerich everchanging resources.