Self-reconfiguration planning (SRP) of modular robots has been a major research for the past decade and still is a key issue. In this paper, we present a novel new graph signature approach for SRP based on Multiple Views Theory, called MVGS (Multi-View Graph Signature). In the proposed method the time complexity of the graph signature generation reduced to O(logn), in which n is the number of modules, from the best solutions with O(n2). Also, we propose a new similarity metric, between graph signatures to better guide the heuristic search toward the final configuration. The proposed approach has been implemented in a simulator and the results show significantly better performance than the other previously proposed methods.