Web mining can be broadly defined as discovery and analysis of useful information from the World Wide Web.The rapid expansion of the Web caused a Constant growth of information which increased the Difficulty of finding relevant information, extracting potentially useful Knowledge and learning about consumers or individual users. In this paper, a fuzzy logic equation is proposed to be used in Web text mining to improve retrieval information from WebPages (HTML files). A comprehensive model of search is presented by including attributes of extracted text from HTML tags they are: title, bold, underline, Font, Font size, center, heading and frequency. These attributes were taken as a measure to compute score for each word extracted. These scores are combined and multiplied by a suggested weight to give a rank to the WebPages.The user takes a query and constructs document vector for it.The similarity measure using a new suggested formula named Content based web retrieval equation is done between the user query and the constructed table stored in database. Experiment results search method based on fuzzy logic significantly improve classifier performance and yields higher accuracy.