Mitochondrial diseases are heterogeneous disorders, caused by mitochondrial dysfunction. Mitochondria are not regulated solely by nuclear genomic DNA but by mitochondrial DNA. It is difficult to develop effective therapies for mitochondrial disease because of the lack of mitochondrial disease models. Mitochondrial myopathy, encephalomypathy, lactic acidosis, and stroke-like episodes (MELAS) is one of the major mitochondrial diseases. The aim of this study was to generate MELAS-specific induced pluripotent stem cells (iPSCs) and to demonstrate that MELAS-iPSCs can be models for mitochondrial disease. We successfully established iPSCs from the primary MELAS-fibroblasts carrying 77.7% of m.3243A>G heteroplasmy. MELAS-iPSC lines ranged from 3.6% to 99.4% of m.3243A>G heteroplasmy levels. The enzymatic activities of mitochondrial respiratory complexes indicated that MELAS-iPSC-derived fibroblasts with high heteroplasmy levels showed a deficiency of complex I activity but MELAS-iPSC-derived fibroblasts with low heteroplasmy levels showed normal complex I activity. Our data indicate that MELAS-iPSCs can be models for MELAS but we should carefully select MELAS-iPSCs with appropriate heteroplasmy levels and respiratory functions for mitochondrial disease modeling.