To deal with the demanding goals of the restricted Diesel emission legislation the focus in the application of exhaust after treatment systems to fulfill these strict requirements moves more and more to model based design, optimization and control. Against this background a nonlinear model predictive control (MPC) for an exhaust after treatment system consisting of a DOC (Diesel oxidation catalyst), a DPF (Diesel particulate filter) and a SCR (selective catalytic reduction) is proposed. Similar to standard ammonia control schemes also in this study the dosing amount is the only adjustable input variable. But in contrast to these usually simple feed forward dosing approaches in this proposal also information about the nonlinear temperature behavior and the storage effects of the SCR -included in a nonlinear SCR model - can be considered for the ammonia control. Based on these improved knowledge of the system a nonlinear model predictive controller, also applicable for real time tests on the engine, is applied for SCR control leading to promising results in simulation.