This　paper　presents　a　harmonic　magnetic　field　detection　technology　for　damage　identification　of　an　in-service　coated　steel　pipeline.　Based　on　the　principles　of　electromagnetic　theory　and　magnetic　field　detection　combined　with　magnetic　focusing　technology,　an　array　consisting　of　a　focusing　detection　probe　and　a　harmonic　magnetic　field　detection　system　were　designed.　However,　the　acquired　detection　signal　includes　an　excitation　signal　in　addition　to　the　defect　information.　In　order　to　make　the　defect　information　more　obvious,　the　excitation　signal　needs　to　be　removed　to　extract　the　defect　feature.　Local　mean　decomposition　(LMD)　is　a　new　time-frequency　analysis　method　that　adaptively　decomposes　a　signal　into　a　set　of　product　function　(PF)　combinations.　The　envelope　of　the　PF　is　the　instantaneous　amplitude　and　the　instantaneous　frequency　can　be　calculated　by　demodulating　the　derivative　of　the　phase　with　a　uniform　amplitude-modulated　signal.　This　method　completely　bypasses　the　Hilbert　transform.　Therefore,　it　does　not　involve　the　problem　of　negative　frequencies　without　physical　meaning.　The　effectiveness　of　LMD　is　demonstrated　by　a　successful　example　of　damage　detection.　Combining　the　data　characteristics　obtained　by　the　experiment,　the　data　processing　algorithm　suitable　for　the　test　data　is　written　by　improving　the　LMD.　The　algorithm　is　easy　to　use　and　has　high　engineering　practicability.