The　damage　will　cause　the　decrease　of　structural　dynamic　property,　such　as　natural　frequency,　mode　shapes　and　damping.　Therefore,　the　structural　damage　may　be　identified　by　the　change　of　dynamic　property.　Most　of　vibration-based　damage　detection　methods　require　the　modal　properties　that　are　obtained　from　measured　signals　through　the　system　identification　techniques.　However,　the　modal　properties　are　not　such　a　good　sensitive　indication　of　structural　damage.　As　the　wavelet　packet　transform　is　a　mathematical　tool　that　has　a　special　advantage　over　the　traditional　FT　in　analyzing　non-stationary　signals,　it　may　analyses　a　signal　in　time-frequency　domain　in　an　arbitrary　resolution.　A　structural　damage　detection　method　based　on　wavelet　energy　distribution　is　proposed　in　this　paper.　Firstly,　the　normalized　vectors　that　present　the　energy　distribution　of　wavelet　packet　components　of　undamaged　and　damaged　structural　responses　are　established　and　named　as　the　parent　vectors　to　reference　model.　Secondly,　the　measured　dynamic　signals　are　decomposed　into　the　wavelet　packet　components　and　the　wavelet　energy　distribution　vectors　(named　as　sample　vector)　are　computed.　Thirdly,　the　damage　are　identified　by　calculating　the　distances　both　sample　vector　and　every　parent　vector.　At　last,　this　damage　identification　method　is　verified　with　a　tested　simply　supported　beam.