World Wide Web (WWW) is a collection of interconnected Web pages accessed via Internet offering information and services from all over the world. There is an explosive growth of information in the Web that poses a challenge to Web users to extract essential knowledge from it. Web Mining is a process of extracting knowledge from WWW using Data mining Techniques. The Web contains Semistructured data like Web pages and Web logs. Mining of Semistructured data involves, (i) extraction of required information from the Deep Web which is not accessible to the search engines, (ii) careful exploitation of extracted data for enhancing the services offered by Web tools such as finding synonyms for search keywords and (iii) ranking the Web pages obtained from the results. To improve the overall performance of the Web, Web logs are mined to derive patterns to predict user behaviour. These patterns can be analysed and used for many applications such as generation of prefetching rules, construction of community Web directories for personalization. These techniques improve availability of web objects, reduce latency and utilize bandwidth efficiently in accessing the Internet.