This　paper　uses　conditional　random　field　model　to　mine　TCM　data.　In　order　to　identify　the　information　of　prescriptions　in　ancient　Chinese　medicine　books,　the　model　firstly　cleaned　and　segmented　the　original　text　data,　then,　labels　are　added　according　to　the　selection　of　TCM　characteristics　to　mark　part　of　the　corpus,　and　according　to　the　demand　divided　into　two　parts,　the　training　corpus　was　used　to　adjust　the　parameters　of　the　custom　CRF　feature　template,　and　the　model　was　generated　after　adjusting　the　parameters.　The　accuracy　of　the　model　for　entity　recognition　was　verified　by　using　the　test　corpus.　In　the　end,　the　feasibility　of　the　conditional　random　field　model　in　the　identification　of　TCM　prescriptions　was　verified　experimentally.　©　Published　under　licence　by　IOP　Publishing　Ltd.