Designing the motion platform for the flight simulator is closely coupled with the particular aircraft’s flight envelope. While in training, the pilot on the motion platform has to experience the same feeling as in the aircraft. That means that flight simulators need to simulate all flight cases and forces acting upon the pilot during flight. Among many existing mechanisms, parallel mechanisms based on the Stewart platform are suitable because they have six degrees of freedom. In this paper, a real coded mixed integer genetic algorithm (RCMIGA) is applied for geometry optimization of the Stewart platform with rotary actuators (6-RUS) to design a mechanism with appropriate physical limitations of workspace and motion performances. The chosen algorithm proved that it can find the best global solution with all imposed constraints. At the same time, the obtained geometry can be manufactured because integer solutions can be mapped to available discrete values. Geometry is defined with a minimum number of parameters that fully define the mechanism with all constraints. These geometric parameters are then optimized to obtain custom-tailored geometry for aircraft flight simulation.