It is generally observed throughout the world that in the last two decades, while the average speed of computers has almost doubled in a span of around eighteen months, the average speed of the network has doubled merely in a span of just eight months! In order to improve the performance, more and more researchers are focusing their research in the field of computers and its related technologies. Internet is one such technology that plays a major role in simplifying the information sharing and retrieval. World Wide Web (WWW) is one such service provided by the Internet. It acts as a medium for sharing of information. As a result, millions of applications run on the Internet and cause increased network traffic and put a great demand on the available network infrastructure. The rapid growth of the WWW and Web development has been the result of many innovative advances in Web technology. Web works with arrays of technologies for better communication with the Internet user, but the inconveniences to users still persistent among the users. A possible solution for this problem is, to add a new resource and distribute the network traffic across one or more resources. Web caching is one such method which is widely used to reduce the network traffic by storing Webpages to a location nearer to the client (Pallis et al 2008). A proxy server is responsible for Web caching which acts as a mediator between the Web server and the Web client and thus reduces latency in retrieving the pages. This proxy-based Web caching system can still be improved to control the performance of the Web. This chapter thus focuses on a methodology for improving the proxy-based Web caching system. It uses Web Usage Mining (WUM) to optimize the performance of the Web based system through Web caching and pre-fetching.