This research aims to solve problems arising from the trust mechanism of multimedia and its mechanism and put forward a Feedback Trust Weighted for Data Fusion algorithm (FTWDF) drawing upon the collaborative filtering algorithm and network fuzzy theory. Also, it has carried out simulation experiments and analyze its performance. It turns out that in the data reliability analysis, the transmission rate of trusted data packets of FTWDF algorithm put forward in this study is higher. In the data precision analysis, it turns out that compared to TMDF and LDTS algorithms, the correct rate under the algorithm put forward in this study is 4.1% and 8.3% higher than TMDF and LDTS. In ML100M and NF5M datasets, the FTWDF-EEFAF model yields a higher precision and thus provides better recommendation results. In the analysis of the number of death nodes, the new clustering algorithm FTWDF-EEFA model serves to increase the survival time of nodes and prolong the lifespan of networks. It improves the survival period of nodes, balances the network load and prolongs the lifespan of networks. In the analysis of energy consumption of nodes, it turns out that the FTWDF-EEFA clustering algorithm can balance the energy consumption of nodes and effectively save the overall energy of nodes. Therefore, through the study, it can be seen that improving existing algorithms serve to effectively increase lifespan of network and improve the trust mechanism. The results are as expected and it offers reference basis for the application of trust mechanism in actual network.