Trust　has　been　widely　used　as　an　effective　means　of　dealing　with　security　issues　in　open　networks.　By　taking　into　consideration　of　the　requirement　on　dynamic　adaptability　and　incorporating　the　factors　of　insufficient　conglomerate　and　incentive　mechanisms　in　trust　evaluation,　we　propose　in　this　paper　a　dynamic　trust　evaluation　model　based　on　multiple　factors　that　is　suitable　for　open　networks.　In　the　proposed　model,　trust　evaluation　takes　into　consideration　of　direct　trust　that　includes　historical　information　to　improve　accuracy　and　recommendation　trust　that　includes　a　bonus-penalty　factor　and　the　reliability　of　evaluation　to　improve　reliability　and　efficiency.　The　calculation　of　integrated　trust　relies　on　solving　the　problem　of　determining　the　weights　between　direct　trust　and　recommendation　trust,　resulting　in　a　balanced　weight　factor　being　introduced　into　trust　evaluation.　In　the　interactions　between　network　entities,　the　balanced　weight　factor　changes　dynamically　as　the　interactions　continue　between　the　entities　to　make　trust　evaluation　dynamically　adaptable.　Experimental　results　show　that　the　proposed　method　can　resist　attacks　from　hostile　entities,　lower　the　influence　of　inaccurate　or　false　recommendations　from　hostile　entities　and　improve　the　accuracy　of　trust　evaluation.