Human mitochondrial genome (mtDNA) variations, such as mtDNA heteroplasmies (the co-existence of mutated and wild-type mtDNA), have received increasing attention in recent years for their clinical relevance to numerous diseases. But large-scale population studies of mtDNA heteroplasmies have been lagging due to the lack of a labor- and cost-effective method. Here, we present a novel human mtDNA sequencing method called STAMP (equencing by argeted mplification of ultiplex robes) for measuring mtDNA heteroplasmies and content in a streamlined workflow. We show that STAMP has high-mapping rates to mtDNA, deep coverage of unique reads and high tolerance to sequencing and polymerase chain reaction errors when applied to human samples. STAMP also has high sensitivity and low false positive rates in identifying artificial mtDNA variants at fractions as low as 0.5% in genomic DNA samples. We further extend STAMP, by including nuclear DNA-targeting probes, to enable assessment of relative mtDNA content in the same assay. The high cost-effectiveness of STAMP, along with the flexibility of using it for measuring various aspects of mtDNA variations, will accelerate the research of mtDNA heteroplasmies and content in large population cohorts, and in the context of human diseases and aging.