The main objective of cell bioengineering is to generate customized tissues that allow recovering the lost functions in the organism in the absence of immune rejection. Although the possibility of in vitro generation of entire organs is technically very complex, obtaining specific cell types for replacement therapies seems to be a more realistic goal at mean time. In this context, those pathologies affected by the dysfunction of a specific cell type, as it is the case of beta-cell in diabetes, would be in principle candidates to benefit from cell transplantation protocols. Embryonic stem cells offer interesting possibilities in this context because they fulfill two important criteria: (i) High proliferation rate by symmetric cell division, overcoming the problem of biomass scarcity and (ii) Plasticity of differentiating to all cell types present in the adult organism, including the germ line. Different approaches have been developed in vitro to obtain insulin-producing cells from embryonic stem cells. Nevertheless, a definitive protocol does not exist yet. However, the experience accumulated in this field by the different laboratories has provided considering key points that would help to design a preferred protocol in the future.