Recognizing　Textual　Entailment　(RTE)　task　is　defined　as　recognizing,　given　two　text　fragments　T　(Text)　and　H　(Hypothesis),　whether　the　meaning　of　H　can　be　inferred　from　T.　This　task　has　broad　application　in　the　field　of　natural　language　processing.　This　paper　discusses　the　classifier　based　method　to　implement　the　entailment　decision.　The　features　are　the　most　important　for　getting　the　correct　result.　The　paper　adopts　different　kinds　of　text　features,　including　the　lexical　feature,　the　syntactic　feature　and　the　semantic　feature.　These　features　are　used　on　different　classifiers　to　verify　their　effectiveness.　The　system　has　been　evaluated　on　the　data　set　of　RTE3-5.　The　best　accuracy　is　60.35%　using　the　SVM　classifier　for　3-way　task.　We　also　explore　the　effect　of　different　features　on　the　system　performance.