In　the　study　and　practice　of　the　tongue　characterization,　experienced　doctors　found　that　a　large　number　of　the　tongue　images　collected　by　tongue　image　instrument　don＇t　meet　the　clinical　requirement,　which　will　directly　affects　the　final　result　of　tongue　image　analysis.　In　this　paper,　the　automatic　quality　evaluation　of　tongue　image　is　designed　for　the　first　time　through　the　following　steps.　First,　the　original　tongue　images　are　processed.　Second,　statistics　of　local　normalized　luminance　based　on　natural　scene　statistics　(NSS)　model,　color,　geometric　and　texture　features　of　tongue　images　are　extracted　respectively.　Finally,　the　Random　Forest　classifier　is　used　to　classify.　Experimental　results　show　that　the　method　we　proposed　can　get　a　better　evaluation　of　tongue　image　quality.　This　approach　can　provide　reliably　reference　data　for　assisted　tongue　image　analysis.