Refactoring involves improve the quality of software and reduce the software complexity without affecting its external behavior. The research focuses code clones is a vital target of refactoring and code clones increase the internal complexity, maintenance effort and reduce the quality of software. A clone consists of two or more segments of code that duplicates with each other on the basis of distinct type of similar measurements. The developed algorithm insist a new semantic based clone management and refactoring system to detect and manage as well as refactor both exact and near-miss code clones. The major goal has to remove the clones in source code fragments by unifying the process of clone detection and refactoring. The implemented clone refactoring technique detects and fixes the clones in multiple classes using graph structure and methods. The code analyzer analyzes the user typed code by separating the auto generated code. Based on a graph structure, a new Abstract Semantic Graph Refactoring algorithm for detecting the clones in multiple classes of source code fragments, have been experimented in this research.