As　an　important　branch　and　application　of　the　Internet　of　Things　(IoT),　the　Internet　of　Vehicles　(IoV)　has　the　characteristics　of　wide　distribution　and　dynamic　connection.　The　current　research　on　trust　measurement　and　management　in　IoV,　to　some　degree,　solved　vehicles　reliability　and　QoS　issues,　but　these　models　still　have　some　drawbacks,　like　insufficient　adaptability　to　the　dynamic　changes　of　the　context.　Therefore,　this　paper　proposes　an　adaptive　trust　measurement　model　for　IoV　based　on　multidimensional　decision-making　attributes.　The　model　not　only　takes　full　advantage　of　the　central　static　trust　management　role　of　the　local　organization　but　also　implements　a　distributed　self-governing　mechanism　to　tackle　the　dynamic　trust　management　issues.　In　the　process　of　trust　management,　the　model　allows　vehicles　to　handle　the　trust　evaluation　according　to　the　service　preferences,　and　vehicles　can　select　some　or　all　of　the　attributes　from　the　multidimensional　trust　decision　attribute　list.　For　the　recommendation　trust　evaluation,　vehicles　can　select　those　vehicles　which　have　similar　service　preferences　from　the　vehicle　candidate　list.　When　computing　the　recommendation　trust,　the　recommendation　trust　dispersion　model　is　used　to　handle　evaluation　bias　problems.　The　method　of　information　entropy　is　introduced　to　tackle　the　weight　adaptation　problem　when　computing　comprehensive　trust　evaluation.　The　simulation　results　and　analysis　show　that　the　model　can　detect　and　recognize　the　malicious　vehicles　in　the　network　and　mitigate　the　risk　that　malicious　vehicles　provide　the　service　to　normal　vehicles.