According to the web information resource distributing features and the requirements semantic-based searching, a semantic recommendation model of information resource by use of publishing/subscribing technologies was presented. The resource is published to the distributing model system by way of message and the users' personalized requests are stored in local inquiring by ways of subscription. The system gives an approach to realize the information services change from users' searching information to server's recommending personalized content, which implements content-based information matching by use of semantic technology. The model system is more efficient deployed than other existing searching educational resource based on web page. Experimental results show that the distributing model system performs better than existent web page recommendation methods in the traditional web page.