This paper investigates multiple neural network-based approaches for voice recognition in computer-based identification systems, with a focus on evaluating and comparing their performance. The methods under examination include artificial neural networks (ANNs), spiking neural networks (SNNs), and a novel identification technique we are currently developing. Each approach will integrate a unique acoustic feature extraction process to enhance recognition precision. Through a series of experiments, we aim to assess the accuracy, reliability, and computational efficiency of each method. Our goal is to determine which model delivers the best balance of performance and feasibility for practical implementation in secure user authentication systems. The system is designed to operate efficiently on personal computers, making it suitable for integration into various platforms, including workplace environments, personal devices, and corporate networks. In addition to testing established ANN and SNN models, we will experiment with hyperparameter tuning and architectural modifications to optimize each network’s performance. Moreover, our novel identification method, which combines advanced neural networks with custom feature extraction techniques, will undergo rigorous testing to evaluate its potential for improving voice recognition accuracy. Initial trials have yielded encouraging results, but extensive experimentation and refinement are necessary to fully validate the effectiveness of these methods. The findings from this study will offer valuable insights into the most promising techniques for reliable voice-based identification, ultimately guiding future development in the field.