Cloning mammals by somatic cell nuclear transfer (SCNT) is highly inefficient. Most SCNT-generated embryos die after implantation because of unidentified, complex epigenetic errors in the process of postimplantation embryonic development. Here we identify the most upstream level of dysfunction leading to impaired development of clones by using RNAi against Xist, a gene responsible for X chromosome inactivation (XCI). A prior injection of Xist-specific siRNA into reconstructed oocytes efficiently corrected SCNT-specific aberrant Xist expression at the morula stage, but failed to do so thereafter at the blastocyst stage. However, we found that shortly after implantation, this aberrant XCI status in cloned embryos had been corrected autonomously in both embryonic and extraembryonic tissues, probably through a newly established XCI control for postimplantation embryos. Embryo transfer experiments revealed that siRNA-treated embryos showed 10 times higher survival than controls as early as embryonic day 5.5 and this high survival persisted until term, resulting in a remarkable improvement in cloning efficiency (12% vs. 1% in controls). Importantly, unlike control clones, these Xist-siRNA clones at birth showed only a limited dysregulation of their gene expression, indicating that correction of Xist expression in preimplantation embryos had a long-term effect on their postnatal normality. Thus, contrary to the general assumption, our results suggest that the fate of cloned embryos is determined almost exclusively before implantation by their XCI status. Furthermore, our strategy provides a promising breakthrough for mammalian SCNT cloning, because RNAi treatment of oocytes is readily applicable to most mammal species.