The growth in the number of anxiety, stress, and depression cases highlights the importance of consistently, objectively, and conveniently tracking emotions and monitoring them. E-textiles are fabrics that incorporate sensors and microelectronics which provide a comfortable and non-obtrusive structural framework of monitoring physiological and emotional conditions when performing everyday tasks. The paper describes an emotion-recognition system that is human-centric and is based on multimodal sensing, machine learning (ML), and real-time biofeedback in soft wearable textile layers. The suggested system will use textile-based electrodes, piezoresistive pressure sensors, capacitive touch sensors, and lightweight electronic modules to record the major indicators of emotions, including heart rate (HR), heart-rate variability (HRV), electrodermal activity (EDA), respiration rate (RR), EEG signals, and micro-pressure patterns. Experimental findings indicate that ensemble classifiers demonstrate the best performance with the highest precision of 99.15, recall of 99, and F1-score of 99.5 as well as textile-integrated EEG sensors can identify early emotion changes at a maximum of 72 percent accuracy. The system additionally offers subtle biofeedback vibration, breathing guided cues and light modulation to help in emotional self-regulation. A textile architecture with multiple layers and all modules is more comfortable, is more durable and can be replaced easily. In general, the present article shows that intelligent e-textile can be used in practice to deliver real-time, accurate, and individualized mental-health monitoring.