With the advance of agricultural modernization, agriculture website was increasingly becoming a major tool for farmers getting information about life and production. How to make the analysis of the needs of farmers effectively to help them to find the information from the ocean of information and resources of the Internet they were interested in had become an urgent and important issue. In this paper, we used the website of Agridata as an example and focused on the solution for the problem. A way was proposed for analyzing and mining the web log of Agridata, which integrated statistical analysis and cluster analysis. By this method, could information behaviors of users be grasped when browsing the website. It was important significance in improving the structure and content of agriculture website, which could provide better services for farmers and improve the level of modernization of agricultural production.