In recent years, Deep Learning (DL) techniques have advanced significantly, enabling the development of sophisticated models capable of addressing diverse challenges. This paper introduces a robust Artificial Neural Network (ANN) model designed for intrusion detection in data networks using network traffic data. Our ANN model is constructed as a fully connected neural network and trained on the Network Security Laboratory - Knowledge Discovery and Data Mining (NSL-KDD) dataset, a widely used benchmark in network security. We explore various network architectures, varying the number of layers and neurons, to optimize performance. Results demonstrate that the ANN architecture featuring a normalization layer as the input, three hidden layers, and an output layer achieves exceptional accuracy, with a validation accuracy of 99.63% and a rapid training time of 4 min and 24 s. Our study contributes to the exploration of DL applications for intrusion detection across diverse data network environments. We also discuss the limitations encountered and outline future research opportunities. The deployment of our proposed model holds significant promise for real-time implementation of DL-based intrusion detection systems, thereby enhancing network security effectively.