Now available in the market for consumer-related goods, including technology, household items, etc., have emphasized personalized , according to the different preferences of each user, planning to conform individual user of different preferences and patterns, so here we import the concept of business intelligence, so that different users according to their usual habits and preferences to browse the Site Map. In this study, we use gray prediction theory to generate a predicted value and the predicted values, using Huffman coding algorithms construct a Huffman tree, using Huffman tree can be ordered out by the Huffman code corresponding, then the use Huffman coding construct of a dynamic website map. In this article we will count the number of times each member browse the Web, and the general number of users browsing the web, a certain weight to change the membership of the exclusive site map.