In the last decades internal combustion engines have become complex mechatronical systems, with a high number of degrees of freedom in terms of actuators, sensors and consequently control tasks. Besides customer requirements the main driving force for the increase in complexity are more and more stringent emission limits imposed by legislation. These limits require a precise control of the combustion process and often an increase of system complexity by requiring new components. While, current engine control applications are still dominated by hierarchical and cascaded schemes, there is a strong interest in optimization based control strategies In this context, advancements in MPC methods have prompted an increased interest in its application in automotive control problems. MPC has evolved from the chemical and process industry and is not directly tailored for automotive applications, because special features of current control strategies and also properties regarding the setup process itself need to be considered. We discuss and combine available tools and propose ways to open gaps towards an optimization based engine control with MPC. To this end, this work combines different approaches and addresses important subcomponents where improvements can gain substantial performance increases, both in terms of achieving targets but also in simplifying the setup process. Our aim is to introduce a seamless framework for MPC based engine control by taking into account the whole system and considering different aspects, such as control system design, sensors, actuators and the tuning itself. This thesis is structured into three major parts, first the search for an optimal reference system for air system control in case of sensor errors, second the utilization of new available cylinder pressure sensors for modeling and control and third the improvement of MPC setup and tuning.