Background: Glycemic variability is suggested to be a predictor for the risk of complications of diabetes. A multitude of parameters to express glycemic variability have been described, but no gold standard exists. The easy measurable parameter SD has been shown to be strongly related to other parameters in a group of patients with mostly well-controlled type 1 and type 2 diabetes mellitus (T1DM and T2DM, respectively). Glycemic variability is higher in T1DM compared with T2DM in mixed populations with different treatments, but studies in patients on intensive insulin treatment are lacking. Therefore in this study we investigate different parameters of glycemic variability and differences between T1DM and T2DM in inadequately controlled patients on intensive insulin treatment. Methods: In this cross-sectional, observational study we describe glycemic variability, measured as SD, coefficient of variation, continuous overall net glycemic action, and mean of daily differences in a cohort of inadequately controlled T1DM (n = 166) and T2DM (n = 58) patients on intensive insulin treatment. Results: SD of 48 h (SD total) was highly correlated to all other measured parameters of glycemic variability (r = 0.66-0.88). All parameters of glycemic variability were significantly higher in T1DM compared to T2DM (P < 0.001), although hemoglobin A1c and mean glucose were comparable and treatment regimen was the same. In the cohort of T2DM patients but not T1DM, a longer duration of insulin therapy was associated with higher glycemic variability. Conclusions: SD total is a conveniently measurable parameter to express glycemic variability in patients with inadequate control with intensive insulin therapy. Patients with T1DM and long-lasting T2DM have the highest glycemic variability.