The objective of this research is to extend the joint parameters identification method developed before to the design of flexible servomechanisms. With the same test data, the parameters identification by means of a genetic algorithm is performed in this paper. The mechanism with five stiffness and six damping coefficients is used for demonstration. Three design variable ranges are assumed in this paper. The results are compared with those of the sensitivity method. The comparison shows that the key parameters with larger sensitivity have less discrepancy between the identified results, less variation, and smaller standard deviations. If tolerances are reduced inadequately for some design parameters, tighter tolerances will be required for the key parameters in order to meet the response requirements.