This paper presents a real-time, wearable-based detection system to identify and distinguish between anxiety and panic attacks using machine learning techniques. Leveraging a Support Vector Machine (SVM) model trained on multimodal physiological data, including heart rate variability (HRV), galvanic skin response (GSR), temperature, and physical activity levels, this system provides an innovative approach to mental health monitoring. The detection system was developed with a goal to enhance real-time classification accuracy, reliability, and applicability, especially in wearable devices like smartwatches, which have become widely accessible. The approach combines real-time data acquisition with robust machine learning algorithms, focusing on SVM due to its efficiency in binary classification problems and capability to handle complex relationships in physiological data. The dataset comprises simulated and augmented data reflecting typical physiological responses for anxiety and panic attacks, facilitating the model’s predictive accuracy across multiple scenarios, including cases where symptoms overlap. A notable aspect of the project is its integration of cross-validation, feature scaling, and hyperparameter tuning. By leveraging Grid Search, the SVM model was optimized to improve prediction accuracy, addressing common challenges in classification models such as data imbalance and feature variance. Comparative analysis with other machine learning models, like Random Forest and Logistic Regression, highlighted SVM’s superior accuracy in this binary classification context. The system also incorporates ensemble learning and hyperparameter tuning to enhance accuracy further. Additional novelty lies in a real-time alert mechanism that notifies caregivers of the detected state, thus contributing to early intervention opportunities for individuals suffering from these conditions. Through simulation and real-world testing, this system achieved a classification accuracy of over 90%, demonstrating its efficacy in distinguishing between panic and anxiety attacks.