In order to understand the differences in the classification characteristics and agronomic traits of millet in Yulin, variation analysis, correlation analysis, principal component analysis and cluster analysis were carried out on the agronomic traits of 12 millet lines in the multi-point regional test in Yulin in 2018. The results showed that the variation coefficient of yield was the most with 29.79%, the variation coefficient of millet grain percentage was the lowest with 5.26%. There was a significant negative correlation between yield and ear diameter, and the correlation coefficient was -0.625, and a significant positive correlation 1000-grain weight, the correlation coefficient was 0.698. The principal component analysis extracted four principal component factors: yield character factor, panicle type factor, plant type character factor and grain yield rate, with a cumulative contribution rate of 91.24%. Cluster analysis results showed that the 12 strains were divided into 3 groups by cluster analysis, and different ways to increase yield were pointed out. The results showed the difference and classification relationship of the representative millet lines in Yulin in the same region, which could provide a theoretical basis for the breeding of high-yield millet in Yulin.