Short term intensive insulin therapy has been reported to induce long term euglycemia remission in patients with newly diagnosed type 2 diabetes mellitus, but the factors that are responsible for long-term remission or hyperglycemia relapse are unknown. Original data of 188 patients with newly diagnosed type 2 diabetes treated with short term intensive insulin therapy was reanalyzed. Patients who maintained glycemic control for 12 months with only life style intervention were defined as remission while those who failed to maintain glycemic control for 12 months as hyperglycemia relapse. Relationships of metabolic control, β cell function and insulin sensitivity with remission time and hyperglycemia relapse were explored. Totally 93 patients achieved 12-month euglycemic remission. Substantial improvement in blood glucose, parameters of β cell function and insulin sensitivity were obtained in both remission and relapse patients. The duration of remission was correlated with fasting plasma glucose measured after cessation of continuous subcutaneous insulin infusion (CSII) therapy (fasting plasma glucose (FPG) after CSII, r= -0.349, p<0.0001). Multivariate logistic regression show that FPG after CSII was independent predictor of hyperglycemic relapse (Odds ratio=1.585, p=0.001). All patients were stratified into three groups according FPG after CSII. As multivariate Cox proportional hazards regression demonstrated, compared with the patients with FPG<6.1mmol/L, risk for hyperglycemia relapse was increased 60% in those with 6.1 mmol/L≤FPG≤7.0 mmol/L (Hazard ratio=1.60, p=0.049), and 1.69 folds in those with FPG>7.0 mmol/L (Hazard ratio=2.69, p<0.0001). Our study demonstrated that fasting plasma glucose after intensive insulin therapy is a convenient and significant predictor for hyperglycemic relapse.