Large　scale　forest　mapping　and　change　detection　plays　a　significant　role　in　the　study　of　global　change,　particularly　in　the　research　of　carbon　source　and　sink.　This　paper　presents　results　from　forest/non-forest　classification　using　ENVISAT-ASAR　data.　Both　pixel-based　and　object-based　classification　method　were　developed　for　ASAR　HH/HV　images　acquired　on　a　single　date.　For　the　object-based　classification,　two　different　strategies　were　proposed:　rule-set　and　threshold-ratio.　Using　as　reference　a　land　use　map　derived　from　Landsat　TM　images　acquired　in　2000,　the　accuracy　of　the　forest/non-forest　map　from　ASAR　AP　data　has　been　found　to　meet　the　requirements　of　mapping　the　Northeast　Chinese　forests　at　large　scale.　©　2011　IEEE.