The emergence of the World Wide Web (Web) technology and the advance of data capturing techniques have lead to exponential growth in amounts of data being stored in Web server logs. This growth in turn has motivated researchers to seek new techniques for the extraction of knowledge implicit or hidden in such data. Designing a web site is a complex problem. Web Server logs provide an opportunity to observe users interacting with the site and make improvements to that site’s structure and presentation. In this paper, we motivate the need for a Dynamic data mining approach for mining user access patterns that uses previous mining results during previous time periods. We present an efficient approach that uses latest results of data mining and new changes in Web server logs to generate new mining rules. The proposed approach is shown to be effective for solving problems related to efficiency of handling data updates and accuracy of data mining results. The proposed approach does not depend on the technique used to generate new frequent user access patterns during the current episode (time period). In our analysis, we have used an Apriori-Like algorithm as a local algorithm to generate frequent user access patterns. The experimental results show that, comparing to Apriori-like techniques, our dynamic approach improves the efficiency of the mining process.