To address the high transmission latency commonly found in location-based routing protocols for the Internet of Vehicles (IoV), a routing protocol for mobile Bayesian and bionic intelligent algorithms is proposed. Firstly, the protocol introduces a dynamic Bayesian network (DBN) on the basis of the original greedy perimeter stateless routing (GPSR) protocol to predict the future position of vehicles, and then determine the farthest and most reliable node from the source node at the next time. Based on this node, a trusted communication range is established, within which the concept of mobility is incorporated to select the next hop. Secondly, if the greedy forwarding condition is not met, causing a switch to perimeter forwarding mode, the protocol uses the shuffled frog leaping algorithm (SFLA), ant colony optimization algorithm (ACO), and left and right hand method to determine the next hop routing path. Experimental results show that, in highway scenarios, the proposed protocol outperforms several benchmark protocols in packet loss rate (PLR) and throughput, including the classical GPSR, ad-hoc on-demand distance vector routing (AODV) protocol, particle swarm optimization based GPSR (PSO-GPSR) protocol, and maximum duration miller GPSR (MM-GPSR) protocol.