Pulse　diagnosis　is　an　important　part　of　the　theoretical　system　of　Traditional　Chinese　Medicine.　In　this　paper　we　propose　a　new　and　an　efficient　framework　to　recognize　pulse　signal　in　nonlinear　angle.　Firstly　the　EEMD　(ensemble　empirical　mode　decomposition)　method　is　used　to　filter　out　baseline　drifting　noise,　and　the　result　is　proved　to　be　effective.　Then　the　MFDFA(multi-fractal　detrended　fluctuation　analysis)　method　is　used　to　get　Hurst　index,　Renyi　index　and　multi-fractal　spectrum.　Hurst　index　is　related　with　the　long-range　correlations,　Renyi　index　is　related　with　the　multi-fractal　characteristics,　and　multi-fractal　spectrum　contains　Δa　and　Δ　characteristics.　Finally,　four　kinds　of　pulse　signals　are　recognized　by　PSO-SVM　after　extract　multi-fractal　spectrum　feature.　Experiment　results　demonstrate　the　effectiveness　of　our　proposed　method.　©　2016　IEEE.