Internet of Vehicle (IoV) is an open network and it changes in constant, where there are large number of entities. Effective way to keep security of data in IoV is to establish a trustworthy mechanism. Through transmission and dissemination of trust, credibility of the entity of IoV is calculated and measured. In this paper a multi-attributes-based trust model is proposed. When the trust relationship between nodes is evaluated, overall experiences of the evaluator are considered as the main reference content, which have a significant restraining effect on malicious behaviors of bad nodes. Moreover, this model combines heuristic algorithm and takes the previous trust evaluation as an important reference content. Thus accuracy of evaluation of trust relationship is improved and sensitivity of this algorithm on behaviors of nodes is enhanced.