Abstract: According to the users' logs into the National Meteorological Service Network, and based on ELK flow log processing technology, the meteorological network log data collection and intelligent analysis system is established. It realizes the analysis of user behavior tracking and data access for different types of users' access to the hot trend analysis and comparison based on a single page retention time, click stream model and the related click index. At the same time, it combines with correlation, clustering analysis and other data mining algorithms to achieve the depth analysis of the site traf fic log. It provides a reference for the construction of more specialized and targeted meteorological data service platform. Through applications in the construction of the National Meteorological Service Network, it is proved that the system is helpful to discover the real needs of the users, and has a positive effect on improving the service ability of the re fined and specialized meteorological data.