Automated driving strategies are capable to improve safety, efficiency and comfort of traffic. To realize their potential, suitable control algorithms are required to steer the vehicle in such a way that the aforementioned goals are reached. Unfortunately, there is high uncertainty in terms of the assessment of the traffic situation but also due to the non deterministic behavior of surrounding human traffic participants. This work investigates a predictive and fuel efficient Advanced Driver Assistance System that controls the longitudinal motion of the vehicle. For this reason, the future motion of the surrounding traffic participants is predicted using data-based, stochastic prediction models. Beside actual measurements of the vehicle?s environment-perception sensors, the utilization of vehicle-to-vehicle and infrastructure-to-vehicle communication is proposed to increase the prediction accuracy. The estimated probability distribution functions of the surrounding vehicle?s future motion are incorporated into a stochastic model predictive control algorithm that computes the fuel-optimal trajectory. In order to improve safety, the prediction uncertainty is considered by imposing chance constraints on a risk function to the surrounding traffic participants. The potential benefits of this approach are demonstrated for an Adaptive Cruise Control application which is evaluated on a single-lane road with traffic lights and on a multi-lane road with frequent lane changes. The evaluations indicate a significant improvement of fuel-efficiency and safety in comparison to non-predictive Adaptive Cruise Control approaches.