Intrusion Detection Systems (IDS) play a critical role in ensuring the security of communication networks and have become indispensable for network administrators. Although numerous methods have been developed for early intrusion detection, many face limitations that reduce their effectiveness against novel or distinct attacks. To address these limitations, this paper introduces the Corporate Hierarchy Optimizer with Multilayer Perceptron (CHO-MLP) approach to strengthen IDS performance. Traditional intrusion detection techniques often encounter challenges such as gradient vanishing, overfitting, and generalization issues. The proposed CHO-MLP framework tackles these problems by leveraging the CHO algorithm to optimize the hyperparameters of the MLP model. This optimization improves the model’s ability to learn efficiently and enhances its detection accuracy. The approach was validated using the NSL-KDD and CICIDS-2017 datasets, ensuring robust testing and evaluation. Using the CHO algorithm, relevant features were selected to refine the classification process, enabling the MLP to distinguish effectively between normal and malicious data. Experimental results demonstrate that the CHO-MLP method significantly reduces training time compared to existing approaches while maintaining high performance across various attack categories. Finally, the efficiency of the CHO-MLP framework was assessed through key performance metrics, including accuracy, precision, recall, and error rate, highlighting its potential as a reliable solution for intrusion detection.