The experiments were conducted for three consecutive years across 14 locations using 9 non-waxy proso millet genotypes and 16 locations using 7 waxy proso millet genotypes in China. The objectives of this study were to analyze yield stability and adaptability of proso millets and to evaluate the discrimination and representativeness of locations by analysis of variance(ANOVA) and genotype and genotype by environment interaction(GGE) biplot methods. Grain yields of proso millet genotypes were significantly influenced by environment(E), genotype(G) and their interaction(G×E)(P<0.1%). G×E interaction effect was six times higher than G effect in non-waxy group and seven times in waxy group. N04-339 in non-waxy and Neimi 6(NM6) in waxy showed higher grain yields and stability compared with other genotypes. Also, Neimi 9(NM9, a non-waxy cultivar) and 90322-2-33(a waxy cultivar) showed higher adaptability in 7 and in 11 locations, respectively. For non-waxy, Dalat, Inner Mongolia(E2) and Wuzhai, Shanxi(E5) were the best sites among all the locations for maximizing the variance among candidate cultivars, and Yanchi, Ningxia(E10) had the best representativeness. Wuzhai, Shanxi(e9) and Yanchi, Ningxia(e14) were the best representative locations, and Baicheng, Jilin(e2) was better discriminating location than others for waxy genotypes. Based on our results, E10 and e14 have enhanced efficiency and accuracy for non-waxy genotypes and waxy genotypes selection, respectively in national regional test of proso millet varieties. The experiments were conducted for three consecutive years across 14 locations using 9 non-waxy proso millet genotypes and 16 locations using 7 waxy proso millet genotypes in China. The objectives of this study were to analyze yield stability and adaptability of proso millets and to evaluate the discrimination and representativeness of locations by analysis of variance (ANOVA) and genotype and genotype by environment interaction (GGE) biplot methods. Grain yields of proso millet genotypes were significantly influenced by environment (E), genotype (G) and their interaction (G × (P <0.1%). G × E interaction effect was six times higher than G effect in non-waxy group and seven times in waxy group. N04-339 in non-waxy and Neimi 6 (NM6) in waxy showed higher grain yields and stability compared with other genotypes. Also, Neimi 9 (NM9, a non-waxy cultivar) and 90322-2-33 (a waxy cultivar) showed higher adaptability in 7 and in 11 locations, respectively. For non-waxy, Dalat, Inner Mongolia (E2) and Wu zhai, Shanxi (E5) were the best sites among all the locations for maximizing the variance among candidate cultivars, and Yanchi, Ningxia (E10) had the best representativeness. Wuzhai, Shanxi (e9) and Yanchi, Ningxia (e14) were the best representative locations, and Baicheng, Jilin (e2) was better discriminating location than others for waxy genotypes. Based on our results, E10 and e14 have enhanced efficiency and accuracy for non-waxy genotypes and waxy genotypes selection, respectively in national regional test of proso millet varieties.