In　order　to　make　full　use　of　multi-resource　information　from　a　structural　health　monitoring　system,　neural　network　and　multi-sensor　data　fusion　technique　were　employed.　to　detect　structural　damage　in　this　paper.　A　5-phase　decision-level　data　fusion　damage　identification　method　that　based　on　the　combination　of　BP　neural　network　and　evidence　theory　was　proposed.　To　validate　the　proposed　method,　six　simulation　damage　patterns　from　a　7-DOF　shear-type　building　model　were　identified.　The　results　show　that　the　proposed　method　cannot　only　improve　the　identification　accuracy　but　also　have　good　adaptive　capability.　This　implies　that　the　proposed　method　is　feasible　and　effective　in　damage　identification.