While　peer-to-peer　networks　bring　convenient　services　to　users,　their　own　characteristics　such　as　openness,　anonymity,　dynamics,　and　autonomy　also　bring　about　some　credibility　issues,　such　as　free-riding　by　malicious　nodes　and　false　evaluations,　malicious　services　and　other　untrustworthy　behaviors.　At　the　same　time,　there　is　a　lack　of　management　mechanism　in　peer-to-peer　network　applications,　and　resource　sharing　between　users　relies　heavily　on　mutual　trust　between　nodes,　which　cannot　ensure　the　authenticity　of　resources.　To　address　this　problem,　the　author　constructed　a　P2P　network　to　simulate　node　transaction　behavior　and　untrustworthy　behaviors　such　as　free　riding　and　malicious　services,　and　obtained　node　dynamic　data.　The　obtained　data　set　is　analyzed　through　the　AHP　analytic　hierarchy　process　to　obtain　the　subjective　weight　of　the　node　attributes,　and　the　principal　component　analysis　method　is　used　to　analyze　the　data　set　to　obtain　the　objective　weight.　The　comprehensive　weight　of　node　attributes　is　obtained　by　weighting　with　game　theory　ideas.　In　this　article,　we　mainly　describe　the　acquisition　of　node　attribute　data　through　simulation　experiments,　and　propose　a　subjective　and　objective　weighting　method　based　on　AHP　and　PCA,　which　effectively　reduces　the　impact　of　user　evaluation　subjectivity　on　weights.