Abstract: To select new sunflower varieties of high-quality and high-yield which are adaptable to Xinjiang''s climate and planting patterns, using 20 sunflower varieties provided by Shanxi Crops Institute as materials, the correlation analysis, grey correlation analysis, regression analysis, and principal component analysis were conducted to explore the relationship between agronomic traits and yield. The results showed that: (1) the variation range of the main agronomic traits of sunflower was 3.41%-17.9%, the variation coefficient of the average flower disk weight was the largest, then was that of the stem diameter, 100-grain weight and disk diameter, followed by that of the leaf number, the variation coefficient of the growth period was the minimum; (2) there was a highly and significantly positive correlation between the yield and seed setting rate, the growth period was significantly and positively correlated with the number of leaves and stem diameter, the plant height had a significantly positive correlation with the stem diameter, the average disk weight and grain width, and the 100-grain weight had a significantly positive correlation with the average disk weight, the grain length and the grain width; (3) the correlation degrees between the traits and yield of the sunflower varieties were: leaf number> growth stage> grain fullness> face diameter and plant height> stem diameter> kernel rate> seed setting rate> grain length> 100 grain weight> grain width; (4) the key agronomic traits affecting the output of sunflower were mainly seed setting rate and kernel rate.