We previously found that Curcuma plants and drugs derived from Curcuma longa, C. phaeocaulis, C. zedoaria, and C. aromatica could be identified by the nucleotide differences at two sites and the existence of a 4-base indel on trnK gene. In this paper, based on species-specific nucleotide sequences, the application of a new method, single-nucleotide polymorphism (SNP) analysis was investigated to identify Curcuma plants more conveniently. First, three types of reverse primer were synthesized in different lengths, 34 mer, 26 mer, and 30 mer, to anneal the template DNAs from each species at sites immediately upstream from substitution positions 177 and 645, and at the site including the 4-base insertion from 728 to 731, respectively. After single-base extension reaction of these primers using fluorescent-labeled ddNTPs and PCR products of the trnK gene region as template, the resulting products were detected using an ABI PRISM 310 Genetic Analyzer. The electrophoretogram showed three or two peaks at different positions depending on the 27 mer, 31 mer, and 35 mer product lengths. Each peak was derived from the incorporated fluorescent-labeled ddNMPs complementary to template nucleotides at positions 645, 724, and 177, respectively. C. phaeocaulis showed three peaks of ddCMP, ddAMP, and ddAMP. The other three species showed two peaks derived from 27 mer and 35 mer products: peaks of ddCMP and ddAMP in C. longa, those of ddCMP and ddTMP in C. zedoaria, and those of ddTMP and ddAMP in C. aromatica. Thus SNP analysis to identify four Curcuma plants was newly developed.