Objectives: This paper aims at clustering documents using a new similarity measure based on energy of a bipartite graph. Methods/Statistical Analysis: We have made use of bipartite representation of documents and clustered them. The proposed algorithm has been illustrated for a small document set. The documents have been clustered using the new similarity measure based on energy of a bipartite graph introduced by us. Findings: Our proposed algorithm gives a better clustering quality comparing with the k means clustering algorithm. Application/Improvements: This proposed algorithm can be further extended and applied to cluster large document sets.