The paper presents nonlinear model predictive control of a doubly fed induction generator (DFIG). The stator power is controlled considering stator voltage dip occurrences as well as rotor speed variation. The nonlinear predictive controller rejects the active and reactive stator power errors by predicting the optimized control voltage over the next sample time and providing reactive power to the system under voltage dip occurrence. The controller performance is demonstrated through computer simulation results using a 11kW DFIG.