The　methods　of　query　expansion　have　been　studied　for　a　long　time-　with　debatable　success　in　many　instances　which　is　one　most　important　technology　to　finish　personalized　information　.　In　this　paper　we　propose　a　probabilistic　query　expansion　based　on　a　new　user　interest　model　which　is　constructed　and　updated　automatically.　We　address　the　two　important　issues　with　query　expansion,　the　selection　and　the　weighting　of　additional　search　terms.　In　constrast　to　earlier　methods,　our　model　deals　with　the　weight　of　those　terms　that　have　not　been　visited　for　a　long　time　which　can　ensure　our　queries　are　expanded　by　those　terms　that　users　are　interested　in　mostly　during　this　period　rather　than　selected　terms　randomly.　Our　experiments　show　that　this　kind　of　query　expansion　has　more　precise　results　than　others.　©2010　IEEE.