The　demand　for　automated　3D　road　bounderies　extraction　is　driven　by　the　importance　of　maintaining　and　updating　the　fundamental　geographic　data　of　road　for　for　various　applications　that　support　urban　planning,　traffic　control,　emergency　response.　Mobile　laser　scanning　(MLS)　as　a　promising　technology　for　the　rapid　3D　mapping　of　road　environment,　provides　a　good　means　to　capture　every　detail　along　the　road　corridor,　including　road　boundaries,　road　markings,　trees,　buildings,　traffic　poles.　This　paper　presents　a　new　automatic　method　to　detect　road　boundaries　in　MLS　point　clouds.　The　geometry　and　intensity　information　are　both　utilized　to　extract　road　boundary　points　from　the　raw　point　clouds.　Experiments　were　undertaken　to　evaluate　the　validity　of　the　proposed　method　based　on　two　test　dataset　captured　by　Optech＇s　Lynx　Mobile　Mapper　System.　The　well　performances　prove　it　to　be　a　promising　solution　for　extracting　3D　road　boundaries　from　MLS　point　clouds.　©　2015　IEEE.