This paper explores the use of neural networks for identifying discrete-time nonlinear systems. The authors develop new parameter estimation algorithms based on a prediction error formulation, derived for neural network models with a single hidden layer. These algorithms are applied to both simulated and real data to demonstrate the effectiveness of the neural network approach. The paper also discusses model validation methods and suggests further research directions, including the extension of the method to more general system models, the comparison with other nonlinear models, and the investigation of different activation functions. The results indicate that neural networks can be an effective tool for modeling nonlinear systems.This paper explores the use of neural networks for identifying discrete-time nonlinear systems. The authors develop new parameter estimation algorithms based on a prediction error formulation, derived for neural network models with a single hidden layer. These algorithms are applied to both simulated and real data to demonstrate the effectiveness of the neural network approach. The paper also discusses model validation methods and suggests further research directions, including the extension of the method to more general system models, the comparison with other nonlinear models, and the investigation of different activation functions. The results indicate that neural networks can be an effective tool for modeling nonlinear systems.