<p>The use of block ciphers is essential in cryptography, ensuring data security by transforming plaintext into unreadable ciphertext. A crucial component of block ciphers is the S-box, which enhances the encryption process by increasing complexity. However, selecting the most suitable S-box for image encryption remains a challenging and uncertain task. To address this, we have developed a novel decision-making model using linguistic neural networks (LNN), specifically designed for S-box selection. In our decision making model, we first gather linguistic term information from three experts. This data is then analyzed using an entropy measure to calculate expert criterion weights. These weights are integrated into the decision process through the Einstein aggregation operator, facilitating the combination of input data with their respective weights. To refine the model, we analyze the hidden layer data using mathematical techniques, particularly using the CRITIC method to compute the hidden layer weights. The output layer aggregates this information, using the Einstein aggregation operator to rank S-boxes based on their security effectiveness. The proposed approach is compared with existing decision making models, and the results demonstrate that our technique is not only applicable but also reliable for decision support in selecting the best S-box for image encryption. The model handles complex decision making scenarios effectively, providing a robust framework for cryptographic applications.. The model improves efficiency while maintaining security standards, making it a superior choice for decision support in image encryption.</p>