We compared effect of long-acting insulin during titration period. Thirty patients with type 2 diabetes were randomly allocated to 3 groups. On admission, fasting plasma glucose (FPG) levels were stabilized at 80 mg/dL level with long-acting insulin, during study period. Group 1: FPG levels were stabilized with glargine 300 U/mL (glargine 300); next, patients wore a continuous glucose monitor device (FreeStyle Libre Pro) and then glycaemic variability (GV) was evaluated on days 3 and 4; glargine 300 was switched to degludec on day 5, and GV was evaluated on days 8 and 9; Degludec was switched to glargine 100 U/mL (glargine 100) on day 10, and GV was evaluated on days 13 and 14. Long-acting insulin was administered at 08:00. Group 2: Long-acting insulin was administered in the order of degludec, glargine 100, glargine 300, following the same regimen. Group 3: Long-acting insulin was administered in the order of glargine 100, glargine 300, degludec, following the same regimen. Area over the glucose curve (<70 mg/dL) was significantly lower in patients on glargine 300, degludec, glargine 100, in that order. glargine 300 may be the best long-acting insulin to reduce hypoglycaemia during titration period.