Nowadays mental health affects people’s lives just as much as physical health and is getting more and more attention, especially in light of the pressures posed by society’s and technology’s rapid advancement. However, diagnosing mental health symptoms primarily depends on skilled psychologists interpreting behaviors and language, which is inaccessible to the majority of people. It has been demonstrated that monotonous speech, prosodic speech abnormalities, and a decrease in verbal activity productivity are all signs of depression. Depression also produces cognitive and physical alterations that impact speech output. The objective of this research is to construct Transfer learning-based model that can screen for mental health difficulties and provide an early diagnostic of mental health problems for individuals. People’s mental health status is captured by an AI-driven model, which focuses on comprehending and analyzing their daily public comments and postings. It does this by analyzing the semantic and syntactic structure of such posts.