In today’s digital world, there is a significant increase in data transfer online especially in multimedia form. This online transmission faces many cyber threats such as denial-of-service attacks, phishing, man-in-the-middle attacks, etc. These threats result in potential security risks and raise the need for secure communication. In this paper, a chaotic map-based encryption technique is presented. The proposed algorithm presents optimized parameter-based 3D chaotic maps for image encryption. The method integrates three primary components: stacked autoencoder (SAE), optimized parameter-based chaotic mapping, and encryption/decryption module, to ensure robust and secure encryption of images. The result evaluated the proposed secure image encryption algorithm with a randomness test, pixel adjacency correlation test, and differential analysis. The mean entropy was approx. 7.9 and the mean number of pixels changing rate (NPCR) was approx. 99.8, unified average changing intensity (UACI) was approx. 33.46. Moreover, the proposed algorithm also investigated the result under noise attacks and shows better cryptanalysis results as compared to comparative state-of-art models.