This study was designed to assess the accuracy with which diabetic patients can estimate their fasting blood sugars (FBS) and to determine whether experience with self-monitoring of blood glucose improves this ability. Twenty patients with type II diabetes who had no experience with self-monitoring of blood glucose were compared with 17 patients who had been monitoring blood sugar regularly for the previous 8 mo. All patients were asked to estimate FBS immediately before it was measured in the laboratory. Patients were very accurate in estimating their FBS; the average error in estimation was 2 mg/dl, and 65% of patients estimated FBS within 20% of actual FBS. However, there was no evidence that experience in self-monitoring of blood glucose improved the accuracy of estimation. Additional studies are needed to determine the types of cues that patients use in estimating blood sugar.