An algorithm is presented to determine the Cole–Cole parameters of electrical impedivity using only measurements of its real part. The algorithm is based on two multi-fold direct inversion methods for the Cole–Cole and Debye equations, respectively, and a genetic algorithm for the optimization of the mean square error between experimental and calculated data. The algorithm has been developed to obtain the Cole–Cole parameters from experimental data, which were used to screen cervical intra-epithelial neoplasia. The proposed algorithm was compared with different numerical integrations of the Kramers–Kronig relation and the result shows that this algorithm is the best. A high immunity to noise was obtained.