In type 1 diabetes field, intrahepatic islet transplantation faces challenges as organ scarcity, poor vascularization, and lack of ECM support at the transplant site. This study introduces a novel human bioartificial pancreas (hBAP) for β-cell replacement using pro-vascularizing cells and ECM support. It aims to replace β-cells by combining endocrine cells, blood-outgrowth endothelial cells (BOECs), and human decellularized amniotic membrane (dHAM). Vascularized islet-like spheroids (SPH) from two protocols (P1 and P2) and native human islets (HI) were compared as endocrine cells for hBAP generation. Same HI equivalents (IEQ) of matched batches were digested with P1 and P2, differing in HI digestion method. HI-derived cells were aggregated alone or with BOEC after 3 days of culture, resulting in Islet SPH and Islet SPH+BOEC respectively. P1 and P2 were compared in term of digestion efficiency, β-cells mass preservation, SPH size and harvesting efficiency. P1 and P2 Islet SPH and Islet SPH+BOEC, alongside batch-matched HI, were characterized through dynamic insulin secretion tests (DIST) and immunofluorescence analysis (IF). Furthermore, hBAP and hBAP(-BOEC) were generated using native HI and cultured ex vivo for 7 days in a customized bioreactor, followed by evaluation using DIST and IF. P1 yielded significantly more total cells after digestion than P2 (p<0.05). P2 showed a significantly higher live cells percentage compared to P1 (p<0.05). Both protocols demonstrated similar efficiency in recovering live cells per IEQ digested. Both Islet SPH and Islet SPH+BOEC from P1 or P2 displayed similar size distributions and insulin secretion performance but reduced insulin secretion capability compared to HI (p≤0.001). Therefore, to standardize hBAP generation, non-digested HI were employed, resulting in hBAP and hBAP(-BOEC) devices. Both bioengineered devices exhibited similar endocrine function to native islets and Ins cells surrounded by a vascular structure. According to the results from P1 and P2, digestion protocol needs to be defined and improved. However, preliminary results on hBAP generated with HI suggested a positive impact of the vascularized dHAM microenvironment in ex vivo function suggesting promising outcomes for in vivo applications. Preliminary in vivo studies are ongoing to evaluate the performance of the hBAP in preclinical diabetic models.