Millions of individuals across all ages, genders, and demographics in the U.S. experience mental and emotional health issues every year. Protracted persistence of negative emotional health significantly influences mental health and can lead to severe mental health illnesses. Traditional methods for assessing emotional health primarily rely on evaluations conducted by mental health professionals through clinical interviews, standardized questionnaires, and physical or neurological examinations. With advancements in deep learning (DL), techniques such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have been increasingly utilized to analyze physiological signals, including Electrocardiogram (ECG), Electroencephalogram (EEG), Electromyogram (EMG), Heart Rate Variability (HRV), and Galvanic Skin Response (GSR), which reflect the body’s response to emotional stimuli. This research aims to develop a high-accuracy CNN model for detecting and classifying emotional states using a publicly available EEG dataset. The proposed CNN classifier is designed to distinguish between negative, neutral, and positive emotional states. To evaluate its performance, key metrics such as accuracy, F1-score, precision, recall, and confusion matrix were analyzed. The model achieved an inference accuracy of 99.77%, demonstrating its potential for enhancing emotional health assessment through AI-driven analysis.