As　a　model　framework　of　outlining　the　network　target　group　and　locking　user＇s　demand　trait,　Personas　can　be　used　to　reveal　the　characteristics　and　behaviors　of　the　user　efficiently　and　accurately.　Setting　the　data　of　SINA　Weibo　users　as　research　object,　a　comprehensive　description　of　the　multi-dimensional　user　characteristics　as　the　main　purpose,　the　K-means　clustering　algorithm　as　a　technical　means,　this　paper　mines　the　user　data　deeply,　then　builds　a　Personas　model,　finally　visualizes　the　results　of　Personas　model　based　on　Matlab.　In　this　paper,　the　proposed　method　of　building　a　Personas　model　provides　a　reference　for　practical　application　problems,　such　as　the　accurate　locking　of　potential　users　and　the　targeted　delivery　of　network　advertising.