In　micro-blogging　services　such　as　the　micro-blog,　the　users　may　get　overwhelmed　by　large　amounts　of　the　data.　An　automatic　method　of　solving　this　problem　is　the　micro-blog　information　classification.　The　traditional　classification　of　microblog　also　lost　some　useful　information　of　micro-blogging.　And　the　weight　calculation　and　feature　selection　of　traditional　method　is　not　an　adaptive　way　to　solve　the　micro-blog　information　classification.　In　order　to　solve　the　problems　above,　this　paper　presents　a　method　of　micro-blog　information　classification　using　the　mixed　characteristics.　On　the　basis　of　the　statistical　methods　of　chi-square　(CHI),　we　use　the　frequency　to　improve　the　feature　selection　and　the　weight　calculation　method.　The　experimental　results　show　that　the　method　is　useful　to　improve　micro-blog　information　classification.　©　2013　IEEE.