Future vehicular networks need multi-hop trusted information among car manoeuvres as a solution to the persistent problem of road safety, and news sharing. However, malicious users in vehicular networks can also disseminate fake information among each other. Traditional public key infrastructure is not an efficient solution for recognising these malicious users, as they all have authorised entities. To cope with this problem, this study highlights novel idea, i.e. three-valued subjective logic (3VSL) as a trust model for multi-hop trust assessment among users in vehicular ad-hoc network (VANET). Trust among vehicle users is represented in the form of opinion derived from 3VSL and updated frequently due to vehicles random movement on the road. To support the authors' proposed scheme, this study contains two parts in simulation, i.e. numerical and experimental analyses. Numerical analysis shows that 3VSL gives accurate trust assessment even with a bridge or random network topology, which is ignored previously by edge splitting. In the experimental part, we extend widely accepted ad-hoc on-demand distance vector routing protocol by directly applying trust fields to the routing table. The simulation experiment shows that their scheme achieves better performance in term of throughput and latencies in low mobility VANET scenario.