To address the limitations of long runtime, weak security, and poor robustness in existing chaos-DNA-based image encryption algorithms, this paper proposes a novel encryption scheme based on a cosine-modulated chaotic map and dynamic DNA coding. A new one-dimensional chaotic map is designed, achieving continuous chaotic behavior across the entire parameter space (\( {{\mu > 0}} \)) with uniform sequence distribution. A dynamic DNA encoding framework is then developed, where eight coding rules, DNA keys, and three operation types (addition, subtraction, XOR) are all dynamically controlled by the chaotic sequence, ensuring global diffusion. Zigzag permutation is further employed to disrupt spatial correlation. Experimental results demonstrate that the proposed algorithm achieves lossless decryption (maximum error 0), cipher image entropy of 7.9957 (approaching the ideal value 8), NPCR of 99.64%, and UACI of 33.73%. The algorithm exhibits strong resistance to brute-force, statistical, differential, cropping, and noise attacks, with encryption time of only 40 ms for 256 × 256 images. These results confirm that the proposed method significantly outperforms existing algorithms in both security and efficiency.
