Characterisics such as non-linearity, uncertainty and time-delays make the control of systems that transition from one operating regime to another very challenging. One means of addressing this problem is to develop a robust transition control structure that is capable of selecting the best controller from a library of possible controllers to put into feedback with the process. This work introduces such a structure for processes that can be represented by a family of single-input/single-output linear models with time-delays and their uncertainties. This framework contains model and controller libraries from which a logic-basic element called a supervisor selects and places into feedback with the process the best controller from the candidate controller library to ensure the desired closed-loop performance of the process. This paper presents the robustness of this transition control structure. The advantage of the proposed transition control structure is that the trade-off between nominal performance and robust stability for a model with large uncertainties is converted into the development of several models with smaller uncertainty, so that the former problem can be solved satisfactorily.