Objective. To detect biomarkers that can be used to predict COVID-19 severity to identify patients with high probability of disease progression and poor prognosis. Methods. Of the 102 patients with confirmed COVID-19 who were admitted to King Fahd General Hospital, Jeddah City, Saudi Arabia, from July 1, 2021 to August 5, 2021, 50 were included in this cross-sectional study to investigate the influence of serum amyloid A (SAA) on disease severity and survival outcomes of COVID-19 patients. Dynamic shifts in SAA, C-reactive protein (CRP), white blood cell (WBC), lymphocytes, neutrophils, biochemical markers, and disease progression were examined. At admission, and at three, five, and seven days after treatment, at least four data samples were collected from all patients, and they underwent clinical status assessments. Results. Critically ill patients showed higher SAA and CRP levels and WBC and neutrophil counts and significantly lower lymphocyte and eosinophil counts compared to the moderately/severely ill patients, especially with regard to disease progression. Similarly, nonsurvivors had higher SAA levels than survivors. The moderately/severely ill patients and the survivors had significantly higher dynamic changes in SAA compared to the critically ill patients and nonsurvivors, respectively, with differences clearly noticed on the fifth and seventh day of treatment. ROC curve analysis revealed that the combination of SAA and CRP was valuable in evaluating the disease progression and prognosis of COVID-19 patients at different time points; however, a combination of SAA and lymphocyte counts was more sensitive for disease severity prediction on admission. The most sensitive parameters for predicting survival on admission were the combination of SAA/WBC and SAA/neutrophil count. Conclusions. The study findings indicate that SAA can be used as a sensitive indicator to assess the degree of disease severity and survival outcomes of COVID-19 patients.