The present study proposes a micromechanical model composed of a two-level homogenization process for predicting the electrical conductivity of polymeric composites incorporating carbon nanotube (CNT) and carbon fiber (CF). The curviness of the CNT and the interfacial resistivity between the polymer matrix and the CNT are considered in the micromechanical model. A series of numerical studies with model parameters are carried out, after which a genetic algorithm is applied to determine the optimized model parameters. The electrical and morphological properties of the composites are experimentally evaluated according to the content of CNT and CF. To verify the predictive capability of the proposed model, the present prediction is compared with that obtained from experiments. The proposed model combined with the genetic algorithm is found to closely simulate the effective electrical conductivity of the composites, as evidenced by the credible accuracy to the experimentally measured values.