As progress in VANETs research continues, there is a persuasive need to support Quality of Service (QoS) routing in such networks. While greedy forwarding is used in many MANETs applications, it is found that it is not convenient for VANETs applications. In this paper, we investigate the important and difficult challenge of QoS routing in VANETs. First, we present an adaptive message routing protocol that uses up to date information about the local topology in order to find the route with minimum end-to-end delay while maintaining a threshold for the connectivity probability and hop count. Then, we propose a genetic algorithm to solve this. To do so, we formulate the QoS routing as a constrained optimization problem. We also derive analytical expressions for the delay as well as the connectivity probability of a route in a two-way street scenario. Numerical and simulation results show that our algorithm gives an optimal or near optimal solutions, which provides an interactive and effective design environment and enriches our protocol performance compared to GPCR.