Web mining treats the World Wide Web as the ultimate data source. This area of research is more interesting than data mining involving extracting data from in-house corporate databases or data warehouses. Gathering information from web sources via distributed intelligent software agents in enterprise architecture can be implemented securely, platform independent, reusable, and would be better modeled in n-tier or enterprise-tier architecture. The purpose of this paper is to develop a novel distributed software agent architecture capable of automatically locating and extracting user specified data from the web, and to dynamically populate semantic databases with the data for later access and retrieval. Implementation of this project not only enables the software to access the semantic information which has been created by a typical web source, but also it is capable of producing a meta database for executing the queries in semantic database. On the other hand, it is too difficult for the current users to interact with distributed intelligent agents over n-tier architecture on the Internet and gather information for web mining issues which have been produced by those agents.