 The paper proposes a new method for detecting chronic obstructive pulmonary disease, COPD, which uses fractional-order dynamics deep learning models to analyze physiological signals such as thoracic breathing effort, respiratory rate, and oxygen saturation. These models were trained using data collected from 54 patients in the Westrow COPD dataset, and 534 patients in the Westrow Porty COPD dataset. The results showed that the fractional-order dynamics deep learning model achieved a COPD prediction accuracy of 98.66%, demonstrating its potential as a reliable alternative to spirometry. Additionally, the model was able to accurately classify COPD stages from stage 0, healthy, to stage 4, very severe. This study suggests that the proposed fractional-order dynamics deep learning model may provide a more accurate and reliable way to diagnose COPD than current methods. This article was authored by Chen Zhongyin, Mihaya Drescu, Gorev Gupta, and others. We are article.tv, links in the description below.