Emotion recognition and prediction plays a vital role in human-computer interaction (HCI), offering more potential for efficient intuitive and adaptive systems. This presents an innovative and efficient approach for emotion prediction from electroencephalogram (EEG) signals by using an Improved Sand Cat Optimization (ISCO) technique to enhance prediction accuracy and efficiency. EEG signals directly indicates the brain activity and these signals are rich and reliable source of data for capturing emotional states. The proposed method is improved by adapting the Cat movement which uses convex lens opposition based learning technique and this will enhance the Cat movement towards target. The proposed method converges to target identification quickly for achieving efficient emotion prediction by extending the exiting Sand Cat Optimization algorithm. The algorithm has been evaluated by using openly available EEG signals dataset, which contains 2132 labelled records of three categories of emotional classes. The performance of the proposed method is compared with other nature inspired optimization algorithms such as Practical Swam Optimization (PSO), Artificial Rabbit Optimization (ARO), Artificial Bee Colony Optimization (ABCO), and Cat Optimization (CO) algorithm. The experimental evaluation shows that the proposed technique outperforms and showcases significant improvements in emotion prediction with accuracy of 97.5% compared to the other bioinspired optimization techniques. This research article has a scope to contribute to the advancement of emotion prediction system in the field of mental health care monitoring, HCI systems, gaming systems, and affective computing.