Mental stress is a normal and frequent phenomenon in human beings. Earlier stress recognition is critical for avoiding these negative impacts since prolonged stress has an adverse impact on mental health leading to anxiety, loss of sleep, or headache. Making utilize of physiological data gathered from a wearable device, present study attempts to simplify the procedure of psychological stress determination, which helps to distinguish the persons suffering from stress over healthy one. We tested our method using a dataset that was made accessible to the public. The precision of forecasting exact stress state was applied to make comparisons among effectiveness of numerous techniques of artificial intelligence (AI), including Artificial Neural Networks (ANN), Fusions of ANN with Support Vector Machines (SVM), Stack Classifying method, and Radial-basis Function (RF) Networks. The study included 3-class stress categorization method in which, results shown greatest accurate rating of 99.920percent by Stack Classifier, while RF provided the lowest preciseness of 84.462percent. Study outcome infer that the suggested models are efficient in detecting mental stress over time and show that physiological indicators could be highly relevant in identifying mental stress.