Mental health disorders are conditions of the mind that cause changes in emotion, thought, and behavior. It can be connected with distress and problems functioning in social, professional, or family activities. This study tries to address this issue and improve the user’s mental health situation by using the proposed LLM-based Agent system. Moreover, this research study tries to determine the user’s current situation through text-based conversation and provide suggestions to improve his mental health problems. Therefore, we developed this system by using LLM-based nCoT (nested Chain of Thought) prompt engineering. In the result section, we sequentially obtained the following: a patient goal achievement score (Average: 5.42), patient Satisfaction score (Mean: 5.57), system response Correctness score (Mean: 6.09), and system Engagement Score (Mean: 6.16). In addition, we analyzed the based CTRS and we obtained all item’s averages (Mean: 5.70), and the standard deviation (Mean: 0.91). Moreover, a system usability scale framework was used for the analysis of the usability (Mean Score: 3.60) of our proposed system.