VR provides a fully immersive experience and can meet the high requirement on controllable variables. The possibility of triggering stronger and more specific emotional reactions in a safer environment arises the interest to use VR as a therapeutic intervention for individuals with mental disorders, yet the use of VR in diagnosis has still not been fully investigated. The main purpose of this project is to create a Dynamic Affective Virtual Environment Gaming (DaveG) that can assist depression diagnosis. Specifically, DaveG would assess user emotions by concurrent psychophysiological monitoring, behavioral analysis, and real-time feedback. It uses different themes, that elicit positive/neutral/negative feelings, to evaluate the emotional reactions. The results of the data analysis will lead to an innovative method of depression detection as a supplementary method to the traditional approaches of diagnosis. Most of the present experiments focus on using VR in exposure therapy that measures the improvement between pre-test and post-test, especially for anxiety disorders. The proposed system is a creative VR-based real-time biofeedback system with self-adaptation to user emotional reactions, which is specifically designed to be used in early depression screening. In future studies, more advanced data analysis methods will be performed to build customized models and serve as the guidance for VR intervention sessions for depression. It also has the potential to be extended to apply to other mental health disorders.