Mining the web data is one of the emerging researches in data mining. The HTML can be used for maintaining the web data but it is hard to achieve the accurate web mining results from HTML documents. The XML documents make more convenient for finding the properties in web mining. Association rule based mining discovers the temporal associations among XML documents. But this kind of data mining is not sufficient to retrieve the properties of every XML document. Finding the properties for set of similar documents is better idea rather than to find the property of a single document. Hence, the key contribution of the work is to find the meaningful clustered based associations by association rule based clustering. Therefore, this paper proposes a hybrid approach which discovers the frequent XML documents by association rule mining and then find the clustering of XML documents by classical k-means algorithm. The proposed approach was tested with real data of Wikipedia. The comparative study and result analysis are discussed in the paper for knowing the importance of the proposed work.