The main objective of this study is to improve the security of data transfer in Internet of Medical Things (IoMT) and also to minimize the processing time and memory size. Methods: The group key, generated by dual encryption scheme(using AES and RSA algorithms), is optimized by proposed Modified Polar Bear Optimization (MPBO) algorithm and then it is used in Rossler Hyper Chaotic System (RHCS) method for medical data encryption and decryption in IoMT. Findings: The total processing times of an image and a signal for their secure transfer by using the proposed Modified Polar Bear Optimization-Rossler Hyper Chaotic System (MPBO-RHCS) are 16.717ms and 6.6593ms respectively. In this case, the memory size of an image and a signal, occupied during their secure transfer are 2.0586 x 1007 bytes and 5.7639 x 1007 bytes respectively. They are the least values of total processing time and memory size of an image when they are compared to other existing optimization algorithms viz. Arithmetic Optimization Algorithm (17.518ms, 3.8851 x 1007bytes), Tunicate Swarm Algorithm (17.72ms, 3.0118 x 1007bytes), Harris Hawks Optimizer (17.004ms, 2.5395 x 1007bytes) and Polar Bear Optimization (17.297ms, 2.3478 x 1007bytes). Similarly, total processing time and memory size of a signal for its secure transfer by using the proposed MPBO-RHCS method are the least values when they are compared to other existing optimization algorithms viz. Arithmetic Optimization Algorithm (7.5129ms, 9.7243 x 1007bytes), Tunicate Swarm Algorithm (7.6142ms, 6.5892 x 1007bytes), Harris Hawks Optimizer (6.9031ms, 6.1497 x 1007bytes) and Polar Bear Optimization (7.0785ms, 6.2333 x 1007bytes).