| Now we are facing the vast sea of network data, to enable the computer to completely and accurately from the quantity big, the structure can be derived from the irregular data is becoming more and more difficult to meet people’s requirement information, before we use search engine to search not only need time is very short, and search the data is very accurate, now we want to better realize the direction of development of information retrieval to the semantic retrieval, but the premise is the semantic annotation, semantic retrieval only the resources on the network to carry on the effective semantic annotation to the semantic retrieval of human dream into a reality.Nowadays, we are faced with vast network data, however, it becomes increasingly difficult to retrieve the required information from these vast and irregularly structured data. Previously we search the required information by search engine for a short time, and the data generated is also accurate, but now if we want to achieve better information retrieval, we need the semantic retrieval. Yet the semantic annotation is the premise of the realization of information retrieval. Only the resources on the network are marked effectively by semantic annotation, can the human dream of semantic retrieval come true.This means that we are now facing such a contradiction between vast date and low retrieval efficiency. Under such contradictory condition, the semantic annotation, especially the automatic semantic annotation gets the favor of researchers, however, because of the different network languages of different countries, the distinctions between different network framework of different Internet ages, and many other reasons, the automated semantic annotation has not yet been fully developed. There are still considerable difficulties need to be overcome to meet the personalized and professional retrieval requirements.Due to such a complex case, this paper first analyzes the principle and method of information collection. Secondly, it makes a full research involving the collective technology related to online national education textual resources, then it manages to obtain online resources through web crawler technology and makes full use of The advantage of domain ontology in semantic expression to express efficiently the Internet data. And thus, the domain ontology base is built to describe the specific subject of relevant network resources. Finally it makes a more penetrating analysis on the Semantic annotation method. By referring to the thinking pattern in soft engineering, taking the Iterative model for the base, and improving the original OCRNIP arithmetic, the model Iterative model of semantic annotation is eventually proposed, and the relevant modeling system design and experimental analysis are finally achieved. |