Mesenchymal stem cells (MSC) are uniquely capable of crossing germinative layers borders (i.e. are able to differentiate towards ectoderm-, mesoderm- and endoderm-derived cytotypes) and are viewed as promising cells for regenerative medicine approaches in several diseases. Type I diabetes therapy should potentially benefit from such differentiated cells: the search for alternatives to organ/islet transplantation strategies via stem cells differentiation is an ongoing task, significant goals having been achieved in most experimental settings (e.g. insulin production and euglycaemia restoration), though caution is still needed to ensure safe and durable effects in vivo. MSC are obtainable in high numbers via ex vivo culture and can be differentiated towards insulin-producing cells (IPC). Moreover, recent reports evidenced that MSC possess immunomodulatory activities (acting on both innate and acquired immunity effectors) which should result in a reduction of the immunogenicity of transplanted cells, thus limiting rejection. Moreover it has been proposed that MSC administration should be used to attenuate the autoimmune processes which lead to the destruction of beta cells. This review illustrates the recent advances made in differentiating human MSC to IPC. In particular, we compare the effectiveness of the differentiation protocols applied, the markers and functional assays used to characterize differentiated progeny, and the in vivo controls. We further speculate on how MSC derived from Wharton's jelly of human umbilical cord may represent a more promising regenerative medicine tool, as recently demonstrated for endoderm-derived organs (as liver) in human subjects, also considering their peculiar immunomodulatory features compared to other MSC populations.