Abstarct: With growing information on the Web, importance of a search engine which enables to provide desired information at low cost and minimum time is vital. An internet user at least once every day uses a search engine to find the required information. Of course, finding a huge-profile list of irrelevant images and results during a search is always a source of inconvenience. One of the most important use of web search engine is extracting existing information on the web, processes them and categorizes information and eventually displays relevant results to the search subject. The result must be reported in order of similarity and relevance to the subject of the search with a particular priority to the user. In this thesis, first different type of search engines are reviewed. Then, we present the challenges for a web search engine. We propose a method using deep learning to extract features from images and texts from web pages to categorize the results for quicker access to desired information.