Abstract: In order to explore the excellent gene resources and broad the genetic basis for wild oats in karst environments. Eighty of wild accessions from different karst districts were investigated by principal component and gray relation analysis. The results showed that, in principal component analysis, the first four main components, which were grains per panicle factor, spike length factor, plant height factor and 1000-grain weight factor, made the contribution of 85% to variation. In gray relation analysis, between plant height and spike length, number of productive panicles and grains per panicle, panicles per plant and plant height, grain weight per plant and grain weight per plant, 1000-grain weight and grain weight per plant were the highest association degrees, therefore, the relationship among them was greater.