Vehicular Ad hoc NETworks (VANETs) are a promising technology that can widely apply to monitor the physical world in urban areas. Efficient data delivery is important in these networks and optimal route selection is vital to improve this factor. Vehicular mobility is a reflection of human social activity and human trajectories show a high degree of temporal and spatial regularity. Therefore, vehicular driving paths are predictable in a large extent. A new opportunistic routing protocol Distributed Partial Order Reduction (DPOR) is proposed in this study that uses driving path predictability and vehicular distribution in its route selection procedure.