Dynamic modeling of synchronous generators is of great importance in power system studies. Three-phase sudden short circuit test is the basic conventional method applied for determining the parameters of synchronous generators. The conventional curve fitting yields to non-unique and inaccurate solutions depending on the initial guesses and nature of the numerical methods. In this paper, a genetic algorithm based method is developed for curve fitting process and identifying uniquely the dynamic parameters through a new fitness function. Owing to the problems associated with sudden short circuit test and unavailability of experimental results, the test is performed on an exact model of the generator with harmonics content space dependent inductances gathered from finite elements analysis. The parameters of a 31.5 kVA alternator are evaluated by use of the proposed method and compared with measured parameters available for this machine by manufacturer showing a good approximation. Copyright © 2009 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.