Abstract-Vehicular networks represent an extremely variable and unpredictable environment. Scenarios can vary from very dense and congested configurations to sparsely populated arrangements. Therefore, protocols designed for such a general scope may fail to efficiently behave in certain configurations. In this paper we propose the Topological-Geographical Routing Protocol (TGRP), a novel solution that presents an adaptive behavior by using a set of standard routing strategies. According to the scenario, TGRP chooses among four different routing approachestwo-hop direct delivery, Dynamic MANET On-demand (DYMO), greedy georouting, and store-carry-andforward technique-to dynamically adapt its behavior to every situation. Performance evaluation shows that TGRP presents a more stable performance under different circumstances, being more adaptable to the changing characteristics of vehicular networks. In fact, TGRP outperforms DYMO by 10% in low density scenarios. In dense networks, TGRP also outperforms the Delay Tolerant Network (DTN) protocol by 10%.