A discrete algorithm for following a desired model for systems that are non-linear or time varying is developed. The controller is developed with optimal control theory by finding the optimal control that would minimize a quadratic cost functional based on the error between the model and system for the previous sampling period. With this information a correction to the system command is formulated for the next sampling period. The stability of the algorithm is investigated and a stability criterion is found. Two examples are presented to show the characteristics of the controller. / Master of Science