Although a growing number of both sequence-based and microsatellite nuclear loci have been used to infer genetic structures, their relative efficiencies remain poorly understood. In our study, we used the Green-backed Tit (<i>Parus monticolus</i>) to explore the resolving ability of these two types of markers. The south-western and central mitochondrial DNA (mtDNA) phylogroups were divergent to some extent in sequence-based nuclear data, while mixed together in microsatellites data. The <i>F</i><sub>ST</sub> values among clades were about four times lower in microsatellite loci than those in sequence-based nuclear loci. We are of the opinion that size homoplasy may have contributed to the inability of microsatellites to uncover differentiation. Our results suggest that sequence-based nuclear loci outperformed microsatellite loci in detecting population structures, especially those focused on populations with large effective population sizes. There was no significant correlation between <i>F</i><sub>ST</sub> values and allelic size variability, which suggested that the efficiency of microsatellite loci in detecting genetic structure may be independent of their polymorphism. <i>F</i><sub>ST</sub> is better than <i>R</i><sub>ST</sub> in detecting intraspecific divergence due to the high variance of <i>R</i><sub>ST</sub>. In agreement with sequence-based nuclear loci, microsatellite loci did resolve the genetic distinctness of the Taiwanese phylogroup. The genetic differentiation between the Taiwanese and continental clades may involve allopatric divergence without gene flow.