This paper presents the design of a new nonlinear expert controller for single-machine infinite-bus and multimachine power systems. The controller uses the concepts of fuzzy linguistic control for normal and abnormal operating conditions of the power system. A nonlinear supervisory controller based on variable-structure system theory is designed to augment the linguistic fuzzy controller for severe transient disturbances when there is a possibility of the stability being lost. This robust nonlinear expert controller is applied to a single-machine infinite-bus system and some other selected machines in a multimachine power system subject to a variety of transient disturbances. The damping performance of this controller is found to be significantly better than the traditional fuzzy and conventional power system stabilizers.