25 years have passed since the first enzymatic isolation of islets of Langerhans from the rat pancreas. During this period of time, it could be demonstrated that transplantation of syngeneic islets intraportally into streptozoticin-treated diabetic recipients (rat, mouse) does not only guarantee long term normoglycemia but inhibits also typical late complications of the disease. Allogeneic islets, however, exhibit strong immunogenicity (rejection within a few days) which can be overcome by various immunomodulating in vitro measures of the islets before transplantation. Isolated islets have been successfully transplanted also in larger animals as dogs and pigs, when transplanted in an autologous or allogeneic system. In human diabetes the success rate of islet transplantation up to now is low and cannot be compared with the results in experimental diabetes. The reasons are manyfold: islet damage due to long ischemia time, low number of islets, low purity, lack of diagnostic markers which indicate rejection, autoimmune destruction. The fact that in the meantime a few patients remain insulin independent after a single islet transplant with the maximum of 2 years indicates however that this method in principle may serve as a tool for the treatment of diabetes. This is underlined by several advantages compared to pancreatic organ transplantation as low risk of the procedure for the recipient and the possibility of repeated transplantation. In addition, in the future xenotransplantation of (porcine) islets might be feasible.