In recent years, there has been growing interest in leveraging physiological signals, such as Electrocardiogram (ECG) data, for emotion classification tasks. This study explores the efficacy of utilizing Transformer models, a state-of-the-art architecture in natural language processing, for emotion classification using ECG signal data. The proposed methodology involves preprocessing the ECG signals, extracting relevant features, and model architecture consists of DistilBERT model, Pooling Layer to obtain a fixed-size representation of the ECG signal, Dropout Layer to prevent overfitting, Fully Connected Layer for classification. Experiments are conducted on publicly available dataset, demonstrating the effectiveness of the proposed approach compared to traditional machine learning methods. The results suggest that DistilBERT Transformer model can effectively capture complex temporal dependencies within ECG signals, thereby achieving notable performance of 76% accuracy in emotion classification tasks. This research contributes to the growing body of literature exploring the intersection of physiological signals and deep learning techniques for affective computing applications.