Image and video (multimedia) databases are very large in size. It becomes very difficult to search the images from these databases by the application of conventional exhaustive searching because it will require unreasonable amount of time. For this purpose an automatic graph based clustering algorithm is developed and proposed through this paper. It reduces the searching time for the images from large databases. The proposed algorithm works on the concept of minimum spanning tree, which removes the inconsistent edges from tree, based on the dynamic threshold provided to the algorithm. The proposed algorithm reduces the search time for the retrieval with an acceptable loss in the accuracy.