In Vehicular Ad-Hoc Networks (VANETs), applications are typically realized by forwarding and exchanging messages among vehicles. The integrity and trustworthiness of messages directly impacts the quality of the applications. Though there have been extensive works on authentication protocols in VANETs, authentication can only certify message origin but cannot guarantee that the identity holders will send truthful and accurate messages. Therefore, in this paper, we propose a novel trust model to directly evaluate the trustworthiness of the content of a message received from other vehicles. The model is built based on various factors such as content similarity, content conflict and route similarity. Trustworthiness of message content will be evaluated at the individual vehicle level and will not introduce any additional architectural assumptions. We have conducted extensive experiments and the results indicate both efficiency and effectiveness of the proposed approach.