Much　prior　work　in　mobile　rover　path　planning　has　been　based　on　assumptions　that　are　unrealistic　for　exploration　of　planetary　terrains.　This　paper　presents　a　new　autonomous　behavior　based　path　planning　approach　for　planetary　rover.　The　constraints　of　rover＇s　maneuverability　and　planetary　terrain＇s　traversability　are　both　satisfied　completely　in　our　approach.　It　works　by　three　steps:　First,　an　expanded　&　colored　obstacle　image　for　path　planning　is　constructed.　Then,　a　serial　of　subgoals　are　planned　using　Path　Finder,　and　a　serial　of　local　optimal　stops　are　obtained.　Last,　a　fine　path　fitting　the　serial　of　substops　is　derived　according　to　rover＇s　maneuverability　and　planetary　terrain＇s　traversability.　At　the　same　time,　we　purpose　the　basic　terms,　definitions　and　theorems.　In　addition,　we　present　the　simulation　results　in　the　real　world.