To　solve　existing　problems　of　Hilbert-Huang　transform(HHT),　such　as　mode　aliasing,　false　mode　and　end　effect　problems　in　the　course　of　empirical　mode　decomposition(EMD),　an　improved　HHT　method　was　proposed　in　this　paper.　At　first,　a　band-pass　filter　method　is　used　to　pre-process　measured　primary　response　signals,　the　sum　of　narrow-band　signals　is　obtained.　Then　a　series　of　Intrinsic　Mode　Functions(IMFs)　are　separated　from　the　processed　signals　by　using　EMD.　The　real　IMFs　are　determined　by　the　correlative　coefficient　between　the　separated　IMFs　and　the　primary　signal.　Afterwards,　the　random　decrement　technique(RDT)　and　Hilbert　Transform(HT)　and　the　least　square　fitting　are　employed　to　identify　structural　modal　parameters　from　the　free　vibration　responses.　The　presented　method　was　used　to　identify　modal　parameters　of　a　measured　steel　frame　finally.　The　results　show　that　the　improved　method　can　extract　mode　parameters　and　transient　features　from　signals　more　effectively,　and　some　problems　existed　in　the　original　HHT　can　be　resolved　successfully.