Railway　track　circuits　is　an　important　device　of　the　train　control　system. Track　circuits　should　realize　the　occupation　detection　of　train　on　the　track.　However,　failure　of　track　circuit　is　unavoidable　and　disruptive,　especially　in　the　main　lines　in　China.　To　realize　the　fault　detection　and　diagnosis　of　track　circuit,　the　paper　proposed　a　hybrid　method　combining　the　model-based　approach　and　the　pattern　recognition　approach.　The　model-based　approach　is　used　to　acquire　the　normal　data　and　faulty　data　based　on　the　multi-track　model.　The　pattern　recognition　approach　is　used　to　extract　fault　features　from　the　simulation　data　and　build　the　fault　diagnosis　system.　In　the　final,　a　test　rig　containing　track　circuit　and　condition　monitoring　system　is　built　to　carry　out　the　fault　diagnosis　test.　According　to　the　test　results,　it　shows　that　the　proposed　method　can　achieve　the　fault　diagnosis　accuracy　up　to　98%　and　fault　diagnosis　time　within　15　s.　©　2022,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.