Abstract: COVID-19, an extremely contagious disease caused by the coronavirus, has led to a global pandemic. Among its prominent symptoms is a persistent cough. Previous studies have suggested that cough sounds carry specific signatures associated with COVID-19. This study highlights the potential of utilizing AI models, specifically deep learning techniques, by collecting a large amount of audio data and divided into training, transfer learning, and testing datasets. The data is pre-processed to ensure a consistent length and transformed into Mel-spectrograms.