Stress is a state of mental or emotional strain due to adversative or challenging situations. A human may undergo bad life experiences or events, and it is a significant issue to be dealt in today's society. It comes as a result of family situations, health concerns, work environments and people's behavior among other things. It affects people of various ages, including children, teenagers, middle-aged people and senior citizens. It has serious consequences affecting human body and life. These problems can be addressed by detecting mental stress at an early stage. The main goal of the propounded system is to identify human stress using a pre-trained Mini-Xception model. Additional measures such as eye moments, mouth activity and head motion are employed for determining the level of human stress. FER2013 dataset is used and the system is tested for all kinds of emotions. Human stress level is experimented by capturing video using webcam. From the experimental outcomes, it is evident that the proposed system offers better performance in terms of Precision, Accuracy, Specificity and Sensitivity in contrast to traditional methods.