<p>Information security is paramount in Internet communications, medical imaging multimedia systems, and military communications. Image cyphering is one of the key techniques to ensure information security during exchanges. We validated our approach on MRI and X-ray images, proving its robustness and effectiveness in encrypting sensitive medical images. This validation confirms the applicability of the hybrid model, combining genetic algorithms and chaotic theories, to ensure enhanced security in medical data exchanges. Thus, this paper presents a new hybrid model based on a genetic algorithm applied at the RNA codon level and exploiting chaotic theories for image cyphering. The model shown is designed in four main steps. First, a transformation of the original image into a binary matrix is performed, followed by DNA encoding via the application of algebraic operators at the nucleotide level. Then, a transition to RNA is carried out, where genetic algorithms intervene using evaluation functions (fitness) applied at the codon level to generate a new population based on chaos theories. In the second phase, a discrimination function divides the image into two types: a high population and a weak population. This segmentation considers the image such as a population, where each line represents an individual. The third phase implements three rounds of an improved Feistel scheme. This process introduces a chaotic coupling between the two categories, relying on a circular motion for the right-hand block and a pseudo-random permutation for the left-hand block. Finally, in the fourth step, an Optimization process is applied using genetic operators, based on entropy criteria, to improve the overall robustness of the encryption. The results of the simulations confirm the quality of the proposed model compared to existing methods. In addition, performance analysis demonstrates the robustness and security of the model in the face of different types of attacks.</p>