Considering　that　the　result　from　single　characteristics　may　differ　from　other　characteristics　and　the　weights　of　aging　characteristics　may　be　not　reasonable　enough　to　judge　the　actual　transformer　oil-paper　insulation　state,　we　put　forward　methods　of　Gray　Clustering　and　Set　Weighting　to　evaluate　the　oil-paper　insulation　state.　First,　we　extracted　the　aging　characteristics　of　return　voltage　method(RVM)　and　polarization　and　depolarzation　current(PDC)　from　the　theoretical　achievements　in　existence.　Second,　we　sorted　out　the　standard　values　of　the　index　of　large　number　of　known　insulated　power　transformer　time-domain　response　data　which　were　used　to　establish　the　insulation　state　classification.　Last,　from　the　collected　data,　we　used　an　entropy　method　and　the　improved　AHP　to　allocate　the　grey　clustering　weights　of　different　aging　characteristics.　The　feasibility　and　accuracy　of　Gray　Clustering　and　Set　Weighting　Methods　in　evaluation　of　transformer　oil-paper　insulation　state　are　verified　by　the　examples　of　field　transformer　test,　which　provides　a　new　direction　for　the　comprehensive　evaluation　of　oil-paper　insulation　state.　©　2018,　High　Voltage　Engineering　Editorial　Department　of　CEPRI.　All　right　reserved.