Continuous subcutaneous insulin infusion (CSII), known as insulin pump therapy, has been widely used for diabetes patients in recent years. Many clinical studies have proved the priority of CSII to multiple insulin injections including better glycaemic control with lower daily insulin requirement, lower glycated haemoglobin (HbA1c) level and reduced risk of severe hypoglycaemia. The best mode of CSII is a kind of programmed and individualized insulin infusion mode. But at present, the applying of insulin pump is mainly during hospitalization period in China, which is not conformed with daily living scenarios of patients. The outpatient insulin pump treatment is much closer to the real living scenarios of patients. However, lack of management experience and widely accepted formative model of insulin pump applying in clinic restricted use of insulin pumps in clinical in China. The investigators aimed to explore a safe and effective management mode of insulin pump operating to enable a wide population to have access to daily use of CSII, and to maximize the rational use of limited medical resources.

In this study, the investigators plan to enroll patients of different types of diabetes, including type 1 diabetes mellitus, type 2 diabetes mellitus, gestational diabetes mellitus, perioperative diabetes mellitus, pancreatogenic diabetes mellitus (after total pancreatectomy) and diabetes patients in perioperative period，each type of diabetes patients were divided into three groups according to the time (7days, 14days and 28days) reaching the target blood glucose. Those patients are educated to record blood glucose and frequency of hypoglycemia during the study period. After achieving goal blood glucose, follow-up will continue one week to evaluate whether the blood glucose reached our goal and adjust the insulin pump parameter if necessary. The investigators plan to apply continuous glucose monitoring(CGM) for each patient during follow up. The investigators aim to explore the best time to reach goal blood glucose considering frequency of hypoglycemia, economic factors, insulin pump parameters at the end of study and the final insulin pump parameters at the end of follow up. The investigators aim to enroll patients with six different types diabetes because those patients have different insulin secretion pattern, which induced to different mode of CSII. The investigators set the goal fasting blood glucose at 4-6mmol/L and the goal postprandial blood glucose at 6-8mmol/L. The blood glucose of hypoglycemia was lower than 3.9mmol/L.

Inclusion Criteria:

1. Patients in clinic of type 1 diabetes mellitus, type 2 diabetes mellitus, gestational/pregestational diabetes mellitus, perioperative diabetes mellitus and patients with diabetes mellitus after pancreatectomy, including those with poor blood glucose control using multiple subcutaneous insulin injections or those are willing to use insulin pump therapy;
2. Age: 18-80 years old;
3. Patients and family members understand the research program and are willing to participate in the study and sign written informed consent.

Exclusion Criteria:

1. allergic to dressings and subcutaneous transfusion tubes;
2. allergic to insulin;
3. Severe diabetic acute complications within 6 months(diabetic ketoacidosis, diabetic hyperglycemia and hyperosmolar state, diabetic lactic acidosis).
4. Severe chronic complications of diabetes mellitus, such as peripheral vascular lesions leading to amputation or chronic foot ulcer, end-stage renal disease, etc. who researchers thought not suitable for this study.
5. Two or more times of severe hypoglycemia occured in the past year (such as consciousness disturbance and coma caused by hypoglycemia), or severe unconscious hypoglycemia.
6. Patients with severe impairment of cardiac function (NYHA grade III or above), liver function (alanine aminotransferase, aspartate aminotransferase or total bilirubin higher than the upper limit of normal value 2 times or more), renal function (serum creatinine higher than the upper limit of normal value) or circulation disorder;
7. Patients with mental illness and self-care ability;
8. Patients or their families could not understand the conditions and objectives of this study.