This study investigates the potential of Virtual Reality (VR) in reducing stress and inducing relaxation, exploring its applicability in trauma and rehabilitation exercises. A calming VR environment featuring nature videos and soothing sounds was developed. ECG signals from 40 participants were recorded under three conditions: resting, artificially induced stress (via mathematical problems, chess games, etc.), and post-VR exposure. HRV metrics were analyzed using machine learning algorithms-SVM, random forest, logistic regression, gradient boosting, and multi-layer perceptron (MLP), with Random Forest achieving the highest accuracy at 58.3 percent. Paired t-tests revealed significant improvements in HRV metrics from the stress phase to the VR phase, indicating VR's effectiveness in stress reduction. Cohen's d effect sizes demonstrated a medium impact of stress on HRV metrics and small to medium improvements post-VR. These findings highlight VR's capability to reduce stress and induce relaxation, supporting its use in trauma therapy and rehabilitation exercises, although the moderate impact suggests the potential need for combined interventions for substantial stress relief.