The　classical　TF-IDF　algorithm　only　considers　the　weight　of　the　term　frequency　and　the　inverse　document　frequency,　without　considering　the　weights　of　other　feature　of　word.　After　the　author　analyzing　summary　of　Chinese　expression　habits,　an　adaptive　weight　of　position　of　word　algorithm　based　on　TF-IDF　is　proposed　in　this　paper,　which　can　be　called　TF-IDF-AP　algorithm.　The　TF-IDF-AP　algorithm　can　dynamically　determine　the　weight　of　position　of　word　according　to　the　position　of　word.　This　paper　introduced　the　vector　space　model　(VSM)　and　designed　comparative　experiment　under　the　scene　of　Chinese　document　clustering.　The　results　show　that　the　F-measure　of　TF-IDF-AP　algorithm　has　been　improved　by　12.9%　comparing　with　the　classical　TF-IDF　algorithm.