Cluster analysis groups objects into a set. The groups are form on the basis of similarities and dissimilarities of objects. Sentence clustering is used in multi-document summarization or text mining. Sentence clustering helps to avoid content overlapping problem. In sentence clustering domains, a sentence is related to more than one theme in document or set of documents. Due to this, the proposed system will be captures fuzzy relationships to increase the scope of problems. In this system, the PageRank and EM algorithm used for sentence level text clustering.