Electroencephalography　(EEG)　recognition　was　one　of　the　key　technology　in　brain-computer　interface　(BCI).　For　motor　imagery　EEG,　a　new　EEG　recognition　algorithm　(DWT-BP　algorithm)　which　combined　discrete　wavelet　transform　(DWT)　with　BP　neural　network　was　presented.　In　DWT-BP,　a　rational　time　window　was　set　through　calculating　the　average　power　of　motor　imagery　EEG　on　electrode　C3　and　C4,　and　then　the　average　power　during　the　time　window　was　taken　into　DWT.　The　combinational　signal　of　approximate　coefficient　A6　on　the　sixth　level　was　selected　as　a　signal　feature　and　BP　neural　network　was　used　as　classifier　to　analyze　the　observed　EEG　data.　The　experiment　results　on　＇BCI　Competition　2003＇　competition　database　showed　that　the　recognition　rate　was　better　than　the　other　several　traditional　algorithms.　So,　it　proved　that　the　algorithm　was　effective　for　EEG　recognition　of　motor　imagery,　and　provided　a　new　idea　for　motor　imagery　recognition　in　brain　computer　interface.　©　2009　IEEE.