Personal identification is a task of authenticating a person using individual biological features. Deep neural networks (DNNs) have demonstrated impressive performance in this field. It is well known, however, that no general algorithm is variable for every application problem. For a new application task, it is very time-consuming for non-experts to design adequately network structure, hyper parameters and an ensemble of base models effectively. In this paper, we present the genetic algorithm (GA) based approach in order to automatically construct network structures, tune their hyper parameters and generate base models for the ensemble algorithm. Then these base models with different network structures are performed to constitute an ensemble. Our original personal identification dataset is employed as the numerical example to illustrate the performance of the proposed method. Convergence property, experiment results, and performance analysis are discussed. The results indicate that several single base models provided by the proposal algorithm have strong classification ability by using the face and voice features and that an ensemble model which is constituted from these base models achieves better classification performance.