In view of such flaws as information redundancy and data loss of vehicle networking system due to a sudden change of vehicle density in urban roads, an adaptive restrictive hybrid routing algorithm (ARHP) based on greedy forwarding (GPSR) and distance vector forwarding (AODV) is proposed to adapt to the rapid change of network environment. As a routing algorithm that combines the greedy forwarding and distance vector forwarding together, ARHP divides the communication network into three modes: greedy forwarding, greedy-distance vector forwarding, and distance vector forwarding. Switching between forwarding modes can be realized based on real-time evaluation of effective node density and network status; meanwhile, a preferable choice can be made of the neighbor nodes by examining the link quality and end-to-end delay performance. Traffic simulator and discrete event simulator are used to compare the performance of ARHP with AODV and GPSR with node density or vehicle speed changed. The results show that ARHP is superior to AODV and GPSR in terms of packet delivery rate and end-to-end delay performance.