Mental Stress has evolved as a major health issue that has negative impacts on humans mentally and physically. It causes significant physiological and psychological changes in the human body. Therefore, continuous observation of mental stress is highly essential for a human which increases the productivity of humans with sound health. A machine learning-based predictive model using Decision Tree (J48) algorithm in cross-validation mode has been proposed to detect the mental stress of any human with respect to different positions of the body and the behaviour of the person. Cost-benefit analysis approach has been proposed and implemented with the machine learning-based predictive model to enhance the accuracy of the model. Due to this, the accuracy of the model has been increased from 80 to 99.67%. Galvanic Skin Response (GSR), temperature, and airflow sensors have been used for the collection of health parameters from humans to have a vast dataset for analysis using machine learning-based predictive model with cost-benefit analysis.