Abstract: A novel Multi-agent System（MAS） framework for edge extraction is proposed due to MAS having parallel operation while the individual agent being simple and easy to implement.The proposed algorithm is based on the description of the diffused gradient on the edge，and is appended some strategies included extending the scope，shifting and mutation of the individual agent for searching the edge.After working together in MAS，agents extract edges effectively.Finally，it is proved in the experiment that the proposed algorithm brings more accurate edge extraction and less sensitivity to the noise.