Abstract:Since the 21st century, much of the knowledge is stored in the document, but the size and structure of all kinds of documents do not facilitate knowledge processing, integration and management. How to extract semantics from the disordered, unstructured data (knowledge) is a task, the primary task is to extract knowledge which is contained in the data and information, to construct the semantic web which is based text resource, using RDF to represent indicate the semantic data, then using TFIDF algorithm to calculates the credibility of the key of the text, the last to enter text information in the database and to realize automatic text classification, ultimate goal is to realize knowledge sharing.