The infection of COVID-19 has caused serious threat to the life and health of all mankind and increased huge economic burden. According to the current statistics, the incidence of pulmonary fibrosis after COVID-19 infection is about 27.7% -87%, 81% of severe patients and 37% of moderate patients have residual lung lesions, and 53% of patients still have residual lung abnormalities one year after infection, resulting in restrictive pulmonary dysfunction and affecting the health and life of patients. Therefore, it is very important to study the diagnostic and prognostic markers of pulmonary fibrosis after infection of COVID-19. At present, relevant studies have been carried out on imagomics and serum proteomics of pulmonary fibrosis after COVID-19 infection, and serum biomarkers and imagomics marker models for diagnosing pulmonary fibrosis after COVID-19 pneumonia have been developed. However, there are few studies combining imageomics and serum proteomics, and the mechanism of pulmonary fibrosis after COVID-19 has not been fully clarified. In this study, it is planned to recruit patients with moderate, severe and critical COVID-19 pneumonia infection, collect venous blood from subjects, and perform chest HRCT follow-up. Blood samples were screened by proteomics and verified by expanded samples to screen diagnostic and prognostic markers of pulmonary fibrosis after COVID-19 infection. At the same time, based on deep learning technology, a model was developed to predict the occurrence and prognosis of pulmonary fibrosis after infection of COVID-19 combined with clinical characteristics, serum markers and AI imagomics, so as to provide ideas for further elucidating the mechanism of occurrence and development of pulmonary fibrosis after infection of COVID-19.