The traditional image encryption algorithm is very simple to be attacked by chosen-plaintext because it has a huge amount of image information data, excessive redundancy, and huge pixel correlation. To upgrade the security of medical images a different DNA coding based medical image encryption algorithm and 3D unified chaotic system is proposed. Initially, a key value is composed from the input image. This newly generated key value is responsible for selecting the chaotic system to be executed and its initial values. Chaotic systems are enforced here to produce a pseudorandom sequence that is utilized for converting the pixels to DNA bases and for scrambling and diffusion operations of plain text images. Finally, simple reversible DNA base conversion rules are used to convert and decode the DNA bases. The decoded cipher image is very hard to recognize without the correct key value. The proposed cryptosystem is simulated, tested, and measured up with some other image encryption algorithms. The outcomes of the several assessments are very favourable and exhibit tremendous encryption performance on different types of medical images, which form the proposed algorithm appropriate for problem solving applications. The experimental impact also prove the durability of the cryptosystem against distinct types of cryptographic attacks.