In the modern fast-paced living environment, people’s psychological pressure problems are increasingly prominent. Biofeedback relaxation training is an effective treatment to relieve stress. It enables users to learn how to adjust their physical and psychological activities to achieve the purpose of anti-stress treatment. Traditional biofeedback training systems mostly use graphs or numerical feedback forms. In order to make the feedback information more vivid and easy to understand, researchers at home and abroad have tried to display biological signals in various forms. In recent years, with the continuous maturity of interface technology, virtual reality technology has gradually been applied to the field of biofeedback relaxation training. However, the current research mainly uses the immersive features of virtual reality to construct a virtual relaxation environment, most of them only have audiovisual stimulation, the interaction and imagination of the environment are rarely discussed. How to integrate more sensory interactions in the environment to further improve the relaxation training effect is still an open problem to be solved.

This study designed a multi-sensory interactive immersive relaxation training application based on the principle of heart rate variability (HRV) biofeedback therapy. The main work includes: collecting HRV data at the user’s fingertips as an indicator to measure the degree of relaxation; controlling changes in the virtual forest environment as a metaphorical visualization of user relaxation; synthesizing guidance voice and sound effects to guide users to respiratory training to improve HRV; Create breathing particles as visual feedback for the user’s respiratory training; add haptic feedback of airflow in the real environment to enhance the user’s perception of “wind speed changes” in the virtual environment. At the same time, I discussed the principles that should be followed when designing virtual reality biofeedback applications from the dimensions of immersion and imagination, multi-modal interaction, virtual-real coupling and virtual-real isomorphism, which provides some new ideas for the design of personal biofeedback relaxation training systems in terms of information display and human-computer interaction.