An autonomous iteration-based approach is proposed to identify faulty links for LEO satellite systems employing intersatellite links (ISL's). A designed network satellite with management responsibilities collects the routing information of communication itself with other satellites, based on which a fault disgnosis model is introduced. Then the most likely faulty link is identified by using an iteration-based probability method, in conjuction withwhich an autonomuos localization procedure is described to pinpoint the actual faulty links. Based on this approach, a hierarchy network management architecture is designed to deal with the fault diagnosis for LEO satellite networks.The performance of the algorithm is evaluated through simulation.First International Conference on Intelligent Networks and Intelligent Systems 978-0-7695-3391-9/08 $25.00