This work presents a powerful application of genetic algorithm (GA) for the identification of Park's model electric parameters of an induction machine. Such a model is used in control techniques for variable speed drives. GA is considered as the most recent product of the artificial intelligence techniques. By its evolutionary character, it solves efficient electrical engineering problems despite its relative slowness in its standard form. Such shortcoming has been dealt with by incorporating a novel adaptive scheme. The suggested adaptive GA aims at accurately solving a nonlinear fitting optimization problem within a reduced computing time. The yielded solution of parameters produces, according to the machine model, the closest possible curves to those of the references. Finally, for the purpose of validation, the obtained machine performances of the adaptive GA method are compared with both those of references and those of a near-least-square-error estimator using experimental variable load measurements.