This paper presents a search-based approach to efficiency optimization of a vector controlled self-excited induction generator (SEIG). The main parts of the control system, besides the SEIG, are a prime mover, a three-phase power converter, and a dc load. The optimization algorithm is executed online and it relies on fuzzy logic (FL) to minimize the internal losses of the SEIG. This is done by adjusting the rotor flux reference. As opposed to existing strategies, additional effort is made to ensure stable operation of the SEIG while the optimization is put on hold and also to avoid the optimization-induced detuning. The iron losses of the SEIG are accounted for in the proposed vector control algorithm as dependent on both the operating flux and frequency. The simulation model of the control system is developed in the MATLAB/Simulink for the purpose of FL controller design, followed by experimental validation. The control algorithm is experimentally implemented using dSpace DS1104 board.