Keyword is useful for a various purposes including labeling, summarizing, indexing, categorization, searching, and clustering. In this paper we will extract keywords from the Arabic text in order to classify it. The proposed system classify any Arabic text through simple statistic and linguistic approaches by extracting the keywords of the text (with their frequency that appear in the text) depending on a Date Base of a particular field (in this work we choose computer science field). This Data Base is represented using one B+ tree for keywords and the other DataBase for non-keywords. The proposed system was implemented using Visual Prolog 5.1, and after testing, it proved to be a valuable for Arabic text classification (From the viewpoint of accuracy and search time).