Mental workload is a critical factor influencing a students’ cognitive performance, especially in complex theoretical courses. It is often associated with stress, cognitive overload, and reduced learning outcomes. With the growing use of AI tools in higher education, understanding their impact on students’ mental workload and emotions has become increasingly important. This research aims to detect and manage mental workload using sentiment analysis, a subset of natural language processing (NLP) to evaluate emotions and opinions. We have included the application of qualitative methodology and surveyed to collect feedback data from students. The study develops hypotheses and introduces related factors like mental stress, physical health, time management, effort, frustration, and engagement to detect mental workload. Sentiment analysis tools such as VADER and TextBlob are then applied to evaluate emotional strain. The findings reveal that the positive sentiment highlighted key aspects of student satisfaction, including instructor support and a manageable workload. Negative sentiment clouds emphasize stress, pressure, and time management challenges. VADER detected more emotion-driven responses, focusing on stress and frustration, while TextBlob captured a more task-oriented view, particularly highlighting academic workload. The research offers useful guidance for developing educational materials. Findings suggest that organized lesson plans and stress-reduction techniques can reduce mental workload. It also provides understanding to create future AI-powered adaptive solutions that improve technological adoption and individualized learning.