Although in a clinical setting islet transplantation is normally performed by percutaneous intrahepatic infusion, the kidney capsule has been the site of choice in nearly all the studies using mice. In the present study, we extensively characterized the mouse model of intraportally transplanted islets with the purpose to propose it as a model to study islet transplantation. C57BL/6 (n = 78) and BALB/C (n = 53) recipients were transplanted with 400 autologous islets alternatively through the portal vein (PV-Tx) or under the kidney capsule (KC-Tx). Glucose concentration during the first hour after syngeneic islet infusion was associated with subsequent long-term function confirming that early events have long-term effects on graft function. In both strains tested the probability to achieve islet function was significantly lower for PV-Tx than KC-Tx. Also in allogeneic models (C57BL/6 to BALB/C, n = 104; BALB/C to C57BL/6, n = 77) the probability to achieve primary function was significantly lower for PV-Tx than KC-Tx and the site of transplantation significantly affected the graft survival. Histological evaluation of livers showed the presence of features (embolism, thrombosis, focal areas of liver necrosis) that are absent in the kidney subcapsular site. Finally, significant differences in the outcome of PV-Tx were observed between the Th type 1 inflammatory-prone C57BL/6 mouse and the type 2 inflammatory-prone BALB/C mouse. Intraportal islet graft model has some features that are more similar to human clinical islet transplantation and should be used as a model to study not only engraftment but also mechanisms of immune suppression and immune tolerance.