In　order　to　study　the　effects　of　alcohol　drinking　on　electroencephalography　(EEG)　during　driving,　a　system　of　indicators　for　EEG　analysis　were　set　up　based　on　the　real-time　EEG　data　of　drivers　collected　by　a　neuroscan　EEG　system　under　virtual　scenes　made　of　sharp　curves,　emergency　stop　belts,　obstacle　zones　and　straight　line　segments.　After　fast　Fourier　transform,　the　mean　powers　of　delta　wave　(δ),　theta　wave　(θ),　alpha　wave　(α),　middle　fast　wave,　beta　wave　(β),　and　gama　wave　(γ)　were　calculated,　and　the　proportion　of　the　power　in　each　frequency　range　to　the　whole　frequency　range　was　obtained.　Then,　the　differences　of　the　combined　indicator　(α+θ)/β　were　compared　between　drunk　driving　and　non-drunk　driving,　and　all　the　indicators　were　analyzed　by　the　paired　sample　t-test.　The　results　show　that　alcohol　drinking　affects　significantly　the　theta　wave,　alpha　wave,　and　gama　wave,　with　t-test　values　being　less　than　0.01;　but　does　not　affect　significantly　the　rest　of　EEGs,　with　t-test　values　being　more　than　0.05.　Among　the　three　parameters　affected　significantly,　only　the　difference　of　gama　wave　before　and　after　drinking　reduces　gradually　with　the　blood　alcohol　concentration　(BAC)　decreasing.　When　the　BAC　dropped　to　40.5　mg/100　ml,　the　difference　of　gama　wave　was　1.12%.