Objective: Glycemic variability (GV) is a well-established and important metric when assessing glycemic control in clinical practice. The objective of this study was to identify factors associated with high GV in patients with diabetes mellitus (DM) hospitalized with COVID-19. Methods: This retrospective, observational study done at three different sites included 685 consecutive patients with DM hospitalized with COVID-19 from March 2020 to December 2020. Demographic characteristics, DM history, comorbidities, laboratory and COVID-19 treatment data were collected. All blood glucose levels, by laboratory serum measurement or by point-of-care testing were collected for the entire hospital stay. GV was expressed as the percentage coefficient of variation for glucose (%CV), derived from the following formula: ([SD of glucose] / [mean glucose]) x 100. High GV was defined as %CV 36%. We used elastic-net regression model (R version 4.1.0, package "glmnet") to select the most important covariates affecting GV of the cohort. Out of the potential 34 variables, 13 had nonzero coefficients and were included in a logistic regression model with high GV as a dependent variable. Results: A total of 685 hospitalized patients were included in the analysis, with a mean age of 67.4±14.1, mean BMI of 32.2±8.3, median (IQR) hemoglobin A1c of 7.4 (6.6-9.0), 323 (47.2%) were male and 425 (62%) were African-American race. The total number of glucose values for the study population was 41,335. A total of 239 (34.9%) patients had %CV of glucose greater than 36%. In the multivariable analysis, the use of systemic steroids [adjusted OR 3.33 (95% CI, 2.21-5.08)], outpatient treatment with insulin [adjusted OR 2.10 (95% CI, 1.42-3.14)], African American race [adjusted OR 1.94 (95% CI, 1.31-2.89)], ICU admission [adjusted OR 1.85 (95% CI,] and CKD greater than stage 3 [adjusted OR 1.81 (95% CI, 1.12-2.94)] were independent factors associated with higher GV. On the other hand, every 5-point increase in BMI was inversely associated with higher GV [adjusted OR 0.80 (95% CI, 0.70-0.90)]. Discussion/Conclusion: Our results show that in hospitalized patients with DM and COVID-19, use of steroids, outpatient treatment with insulin, African-American race, ICU admission and CKD greater than stage 3 were independently associated with higher GV. Otherwise, BMI was inversely associated with higher GV.