When it comes to medical image registration, meta-heuristic algorithms are lifesavers. They tackle optimization problems and help diagnose a wide range of ailments, including malignancies like brain, lung, and breast. Whether it’s multi- or mono-modal registration, intensity-based methods use image alteration to merge many images with similar content into a single representation. Improving the similarity metric between images are of utmost importance. This work modifies the teaching and learning phases procedures in the Modified Teaching-Learning-Based Optimization (mTLBO) algorithm to boost the search space disruption potential. Outperforming the mTLBO algorithm, a self-learning method is introduced to improve the learner’s ability to innovate and global exploration performance. To register COVID-19 lung Computer Tomography (CT) scans, we provide a mTLBO algorithm. The suggested technique outperforms than the Particle Swarm Optimization (PSO) algorithm in terms of registration accuracy and resilience, as shown in the simulation results.