Given the difficulty that is normally associated with tuning highly nonlinear control systems, such as seen in LEO spacecraft, it is interesting to evaluate how new controller types perform. The measured data from the GOCE mission has led to the creation of high fidelity torque models describing the angular perturbations acting on a spacecraft. These torque models present an excellent opportunity to test the performance of new magnetically actuated attitude control methods. The work done in this thesis used these torque models in a simulation environment to evaluate the performance of an INDI-based control architecture for controlling the attitude of a GOCE-type spacecraft, using magnetic control actuation only. The goal was to evaluate whether an INDI-based control architecture can meet the same requirements that were set for the GOCE mission, and simultaneously evaluate the difficulty of tuning the controller and its sensitivity to errors and changes in gains once it is tuned. To achieve this a PD-controller was used as a reference controller, and two INDI-based controllers were set-up. One with a linear control outer loop (LINDI), and one with an NDI outer loop (NINDI). All three controllers have been tuned using a genetic algorithm.