In the paper, a few new results are presented relating to model reference control (MRC) and model reference adaptive control (MRAC) for nonlinear uncertain plants of a very general nature. Noa prioriknowledge about the structure of uncertainties (plant terms and coefficients) is assumed, nor is any constant bound on their values. Such plant models result from many applications. In particular, if uncertainties represent some unmodelled dynamic effects, they will depend on the system states and so be unbounded. The Lyapunov type of stability is considered. A new MRAC law has been developed, which provides stable tracking to within a certain comput able tolerance. The paper includes a few examples of the application of the developed adaptive con troller to nonlinear plants with potentially unbounded terms. The tracking control of robotic manipulators has been especially elaborated.