Skeletal remains encountered frequently in forensic applications are a challenging specimen, since their DNA is usually degraded due to harsh conditions, limiting the utilization of skeletal DNA. Forensic scientists have tried various methods to extract DNA from skeletal remains of low quantity and poor quality or improve detecting technology for more information from compromised DNA. Compared with traditional capillary electrophoresis (CE), massively parallel sequencing (MPS) is more sensitive to shorter fragments, able to detect allele sequences for variations from core motif or flanking regions, and able to detect more markers with a higher discrimination power. In this study, short tandem repeats (STR) and single nucleotide polymorphisms (SNP) from 35 human skeletons were genotyped by MPS platform, and CE method was also used to perform STR genotyping. The results indicated that the detection rates reached 100.00% in 16 of 35 samples with MPS method, while the same 100.00% was reached in only 9 samples with CE. The success rates of MPS were also higher than that of CE method in shared 21 loci (excluding Y-indel, DYS391, and SE33), especially in loci detected by MPS method only. Besides, all SNPs (124 and 90 SNPs in males and females) were detected in 18 samples of 35 samples by MPS method. Some intra-allelic sequence variants were observed in eight loci (D21S11, D8S1179, D5S2800, D3S1358, vWA, D2S1338, D1S1656, D12S391) using MPS technology. Interestingly, there is a sample showing genotyping disagreement in FGA locus. The clone sequencing verified that a "T" deletion discovered in flanking sequence of FGA led to wrong genotyping on Ampliseq Converge. Our results indicated that MPS could be adopted in qualified labs as a supplementary when the DNA of skeletal remains are hard to identify.