Tethered Segway is a robotic platform inspired by human climbers. It is a two-wheeled mobile platform tethered to the top of a structure in order to climb steep surfaces with varying slopes, such as domes. The unstructured environment may cause uncertainties in the dynamic behavior of the robot while operating on different parts of the dome. In this paper analysis and synthesis of a robust controller for a tethered Segway is presented in order to provide desired performance in the presence of uncertainties. To design the robust controller, structured and unstructured uncertainties of the model are encapsulated into a structured singular perturbation. A linear robust controller is designed such that the robust stability of the closed loop system is preserved in the presence of modeling uncertainties. Finally, the effectiveness of the proposed controller is verified through simulation by comparing its closed loop transient response and sufficiently suitable steady-state performance to that of a previously proposed LQR controller for the robot.