Generation of surrogate β-cells is the need of the day to compensate the short supply of islets for transplantation to diabetic patients requiring daily shots of insulin. Over the years several sources of stem cells have been claimed to cater to the need of insulin producing cells. These include human embryonic stem cells, induced pluripotent stem cells, human perinatal tissues such as amnion, placenta, umbilical cord and postnatal tissues involving adipose tissue, bone marrow, blood monocytes, cord blood, dental pulp, endometrium, liver, labia minora dermis-derived fibroblasts and pancreas. Despite the availability of such heterogonous sources, there is no substantial breakthrough in selecting and implementing an ideal source for generating large number of stable insulin producing cells. Although the progress in derivation of β-cell like cells from embryonic stem cells has taken a greater leap, their application is limited due to controversy surrounding the destruction of human embryo and immune rejection. Since multipotent mesenchymal stromal cells are free of ethical and immunological complications, they could provide unprecedented opportunity as starting material to derive insulin secreting cells. The main focus of this review is to discuss the merits and demerits of MSCs obtained from human peri- and post-natal tissue sources to yield abundant glucose responsive insulin producing cells as ideal candidates for prospective stem cell therapy to treat diabetes.