Abstract
Stemming is an attempt to reduce word to its root form. It is a pre-processingstep in Text Mining Applications as well as a prevalent need thing of Natural Language Processing (NLP) filed, Information Retrieval systems and text classifiers. Many research focusing on extracting Arabic root to reduce different grammatical forms of words like its nouns, verbs, and adjectives. In this study we developed a new Text Mining model consist of two algorithms, the first is a stemming algorithm which process the word and match it with the suitable diacritics or vocalize word pattern as (sic) f-a-l and (sic) facil then extract root without using root file. When we applied our Stemmer algorithm on (Sahih Al-Bukhari) textbook, which is one of seven primary books in Al-hadith Al-Sharif in Prophet's Sunnah beside Muslim and others textbook, we achieved 95.8% accuracy of root extraction and 96.4% inflection accuracy. The second algorithm is concerned to mining the knowledge by detecting the entities as verb time (Past, Present, and Imperative), noun and proper noun.