With　the　growth　of　micro-blog　users,　the　number　of　micro-blog　text　is　also　showing　an　explosive　growth　trend.　Faced　with　such　a　large　amount　of　text　data,　how　to　effectively　retrieve　useful　information　is　very　important　for　micro-blog　users.　This　paper　proposes　a　method　combining　traditional　TF-IDF　computing　and　LDA　topic　model.　First,　we　compute　by　TF-IDF　to　find　micro-blog　articles　about　word　frequency　similarity.　Then　we　use　the　LDA　topic　model　approach　to　filter　out　micro-blog　articles　with　similar　themes.　Experimental　results　show　that　using　the　integrated　search　method,　users　can　retrieve　more　suitable　user＇s　actual　needs　micro-blog　articles.