The Nonlinear Generalized Minimum Variance (NGMV) controller introduced in this paper is an optimal control strategy for a class of nonlinear plants. It is assumed that the plant model can be decomposed into a set of delay terms, a stable input nonlinear subsystem and a linear subsystem which can be represented in either polynomial or state space forms. The optimal controller is obtained by minimizing the variance of the generalized signal consisted of error weighting, state weighting and control signal weighting. The controller is easy to compute and the resulting nominal closed loop system is stable. The controller can be represented in the Smith predictor form while the system is open loop stable. In this paper,the development of Nonlinear Generalized Minimum Variance control theory is reviewed in details. Next, NGMV controller design methods for the systems in both state space form and polynomial form are introduced. Finally, future work directions as well as challenging research problems in this area are addressed.