Assessment　and　identification　of　human　activity　and　posture　with　triaxial　accelerometer　can　provide　information　about　the　health　state　and　the　chronic　care.　This　paper　proposes　a　human　state　recognition　algorithm　based　on　Kalman　filter　(SRKF),　which　could　identify　steady　state　and　state　transition　in　real　time.　In　this　study,　an　automatic　state-recognition　system　consisting　of　a　Bluetooth　module　and　a　smart　phone　is　developed.　The　Bluetooth　module　with　a　triaxial　accelerometer　was　placed　on　the　body　and　the　raw　sensor　data　was　transported　to　and　processed　on　the　smart　phone.　Associated　kinematics　characteristic　of　human　activity　and　the　accelerometer　signal,　the　results　of　amplitude　function　of　vector　CSVM　were　processed　by　Kalman　filter　to　identify　human　state.　Data　were　collected　from　ten　adults　in　unsupervised　environment.　Experiment　result　shows　that　the　algorithm　has　achieved　better　performance　on　the　smart　phone　with　limited　computing　and　storage　capacity.　Copyright　©　2012　ICST.