Self-monitoring of blood glucose (SMBG) is well acknowledged in type 1 diabetes mellitus (DM). However, although its use is widely recommended in some developed countries, with a significant economical cost, SMBG remains questionable in type 2 DM because its efficacy is not demonstrated by high-quality randomized controlled trials. Type 2 diabetes is an heterogenous disease and its natural history covers several decades. Although type 2 DM represents an important risk of microvascular and cardiovascular complications, it remains asymptomatic for a long time with often a very poor metabolic control and low drug compliance. Treatment of type 2 diabetes varies during these decades, changes in lifestyle at the very beginning, then using oral drugs without risk to develop hypoglycemia (metformin, glitazones, alpha-glucosidases inhibitors) and/or insulin secreting drugs (sulfonylureas or glinides) with potential hypoglycemic risks. At every stage, SMBG could sensitize the patient and possibly could improve glycemias and compliance. SMBG must be adapted to provide informations representing a very useful, motivating and pedagogic tool. Obviously SMBG, in non-insulin-treated type 2 DM, targets other objectives when compared to its use in type 1 DM requiring adapted educational programs to avoid its uselessness and depression of the patient. In insulin treated type 2 DM patients, with a bed time regimen, SMBG represents both a major tool for adjusting the insulin and oral hypoglycemic drugs doses and a comprehensive instrument for controlling and understanding the disease. Thus SMBG is a logical tool for the management of a large proportion of type 2 DM patients but it requires to be proposed in structured counseling educational programs adapted to the psychological and social profile of the patients. These programs must be evaluated by randomized controlled trials.