A barometer for the emotional, cognitive, and social well-being of a human being is mental health assessment carried out by Clinicians, where how a person thinks, feels, and handles situations are instructed concisely. An element crucial at every phase of life is mental health right from childhood to undergoing adolescence and up to adulthood. Here we are trying to implement a system that will predict the different stages of mental illness in the early-stage for juveniles (10 to 21 years). Category of data have been used to detect mental illness at the onset of the various techniques using massive data where Machine-learning algorithms are used for predicting mental illness. The feedback received from the juveniles through the questionnaire was first subjected to unsupervised learning techniques. The proposed model provides the functionality to determine the mental health of the juvenile with minimal resources required to cover the maximum scope of urban and rural areas. The analysis can be of great benefit to the analytic result, examinations, and prediction of the mental clinic to improve the victim condition and direction of future work also.