Objective To establish a rabbit model of hypoglycemia and evaluate the accuracy and timeliness of hypoglycemia monitoring by continuous glucose monitoring system (CGMS). Methods Sixteen female New Zealand white rabbits were randomly divided into 4 groups, with 4 rabbits in each group. The rabbits in the control group were given intravenous infusion of saline. The animals in the experimental group were infused with insulin continuously, which were divided into 0.1 U/kg/h insulin group (RI=0.1 U group), 0.2 U/kg/h insulin (RI=0.2 U group) group and 0.4 U/kg/h insulin group (RI=0.4 U group) accordingly. During the experiment, CGMS was monitored for 240 min. Blood samples were collected at a 30-minute interval and the blood glucose level was measured by a hand glucose meter. Results A total of 1296 CGMS monitoring data were obtained during the study period, and 136 BG monitoring data matched with CGMS time were obtained. After the insulin administration, BG and CGMS were significantly decreased. The reduction rates of BG and CGMS were 0.016 and 0.017 mmol/L/min in the RI=0.1 U insulin group, 0.04 and 0.027 mmol/L/min in the RI=0.2U insulin group, and 0.049 and 0.032 mmol/L/min in the RI=0.4 U group. According to whether BG monitoring value was lower than 4.4 mmol/L, the BG-CGMS paring data were divided into hypoglycemia and normoglycemia. In hypoglycemia, the average deviation of BG-CGMS was 0.55 mmol/L (the upper and lower limits were -0.98 and 2.08 mmol/L, respectively) and the absolute difference percentage (RAD) was 40.2% ±45.2%. The mean deviation of BG-CGMS in normal blood glucose was -0.19 mmol/L (upper and lower limits were -1.38 and 1.00 mmol/L, respectively) and 5.8% ±5.3% in RAD. The error grid analysis (EGA) showed that the proportion of zone A was 93.4%, 0.7% in zone B, and 5.9% in zone D, and the zone D was distributed in area of low BG and high CGMS. Conclusions The results of this study indicate that CGMS has a significant hysteresis phenomenon when blood glucose is reduced rapidly. When the blood glucose levels fall below 4.4 mmol/L, CGMS may have a risk of overestimating blood glucose. Such risk should be fully considered during CGMS clinical application.