Aiming　at　the　problem　that　oil-paper　insulation　aging　evaluation　accuracy　is　vulnerable　to　single　characteristic　parameter　under　different　circumstance,　a　new　comprehensive　gray　evaluation　method　for　transformer　oil-paper　insulation　status　based　on　dimension　reduction　techniques　and　K-means　clustering　is　proposed.　First,　a　terminal　two-point　analytical　method　of　depolarization　current　spectrum　is　presented　to　extract　three　new　features　of　insulation　aging.　Second,　a　multivariate　evaluation　index　array　is　constructed　and　the　comprehensive　index　Z　is　refined　via　a　combined　weighting　method　combining　with　main　component　dimension　reduction　techniques.　Finally,　a　hierarchical　database　of　＇good,　middle,　and　poor＇　3-level　for　Z　is　built　by　weighted　Euclidean　distance-based　K-means　clustering　algorithm,　the　level　limit　standard　table　for　comprehensive　index　is　determined　and　the　grey　clustering　evaluation　system　for　transformer　oil-paper　insulation　aging　state　is　set　up　accordingly.　The　final　numerous　case　analysis　proves　that　the　proposed　method　is　comprehensive　and　accurate　to　evaluate　the　insulation　state.　©　2019,　Power　System　Protection　and　Control　Press.　All　right　reserved.