Aims/introduction: The predictive low glucose management (PLGM) system was introduced in March 2018 in Japan. Although there are some reports demonstrating the benefit of PLGM in preventing hypoglycemia, no data are currently available in Japanese patients with type 1 diabetes mellitus (T1DM). The aim of the present study is to evaluate the effect of PLGM with sensor-augmented pump therapy in the prevention of hypoglycemia in Japanese patients. Materials and methods: We included 16 patients with T1DM who used the MiniMed®640G system after switching from the MiniMed®620G system. We retrospectively analysed the data of the continuous glucose monitoring system in 1 month after switching to MiniMed®640G. Results: The area under the curve (AUC) of hypoglycemia of < 70 mg/dL was lowered from 0.42 ± 0.43 mg/dL day to 0.18 ± 0.18 mg/dL day (P = 0.012). Correspondingly, the duration of severe hypoglycemia (< 54 mg/dL) was reduced significantly from 15.3 ± 21.7 min/day to 4.8 ± 6.9 min/day (P = 0.019). The duration of hypoglycemia was reduced, but the reduction was not significant. Regarding the AUC for hyperglycemia > 180 mg/dL and the duration of hyperglycemia did not change. With the PLGM function, 79.3% of the predicted hypoglycemic events were avoided. Conclusions: The hypoglycemia avoidance rate was comparable to those in previous reports. In addition, we demonstrated that PLGM can markedly suppress severe hypoglycemia without deteriorating glycemic control in Japanese T1DM patients. It is necessary to further investigate the effective use of the PLGM feature such as establishing a lower limit and the timing of resumption.