With　the　rapid　development　of　the　Internet　and　technological　innovation,　the　number　of　computers　and　resources　connected　to　P2P　(Peer　to　Peer)　networks　is　increasing.　It　has　become　a　trend　to　share　all　kinds　of　hardware　and　software　resources　through　the　network　to　provide　unified　and　open　computing　and　resource　interaction　environment.　Peer-to-Peer　networks　have　the　characteristics　of　heterogeneous,　dynamic,　and　distributed,　which　make　it　challenging　to　provide　trusted　and　secure　application　environments　and　services　for　users　in　this　scenario.　Trust　management　system　is　a　supplement　to　traditional　security　technology　and　is　an　important　way　to　realize　trusted　networks　and　trusted　computing.　While　the　unconstrained　behaviors　of　users　and　the　characteristics　of　decentralization,　scalability,　anonymity,　and　autonomy　make　it　difficult　to　ensure　the　trustworthiness　of　the　application　environment　and　vulnerable　to　malicious　attacks.　To　resolve　these　problems,　researchers　have　proposed　reputation-based　trust　models,　obtaining　inspiration　from　the　establishment　of　trust　relationships　in　human　society.　However,　the　aggregate　calculation　methods　of　reputation　degree　are　subjective　and　vulnerable　to　malicious　behaviors　based　on　the　historical　interactive　and　feedback　evaluation　information.　In　this　paper,　we　are　proposing　a　behavior-based　trust　management　model　for　P2P　networking　applications,　where　we　base　our　method　on　＇Consistent　with　words　and　deeds＇,　the　evaluation　of　behavior　credibility　using　fuzzy　sets,　and　create　the　trust　calculation　not　restricted　to　feedback　trust　which　is　aggregated　by　subjective　service　feedbacks.　Our　simulation　experiments　indicate　that　our　proposed　model　could　maintain　a　good　performance　under　the　malicious　services　of　malicious　nodes.　©　2022　IEEE.