This　paper　proposed　a　damage　identification　method　using　the　kernel　principal　component　analysis　(KPCA)　and　the　support　vector　machine　(SVM).　Firstly,　kernel　parameters　of　KPCA　are　optimized　by　particle　swarm　optimization　(PSO),　then　the　optimized　KPCA　is　used　to　extract　non-linear　features,　and　finally　a　SVM　model　is　regarded　as　the　classifier.　A　numerical　model　of　a　reinforced　concrete　frame　with　12　storeys　was　analyzed　in　order　to　validate　the　method.　The　results　indicated　that　the　proposed　method　not　only　can　effectively　extract　nonlinear　characteristics　and　reduce　data　dimension,　but　also　has　high　capabilities　of　damage　identification,　anti-noise,　generalization　and　strong　robustness.　©Civil-Comp　Press,　2011.