Abstract: [Objective] To identify the key agronomic traits that greatly influence the yield of Sorghum-Sudan hybrid grass, unlock its high-yield potential, and lay the foundation for variety breeding, promotion, and application. [Method] The agronomic traits of 12 materials of Sorghum-Sudan hybrid grass were measured. The correlation, principal component, and path analysis between main agronomic traits and dry weight per plant were performed using a random block test. [Result] The coefficient of variation of the different agronomic traits ranged from 11.61% to 54.67%, showing that the principal agronomic traits varied widely across the materials evaluated. The correlation between the evaluated agronomic traits and dry weight per plant were plant height > stem diameter > number of leaves > leaf length > leaf width. The strongest correlation coefficient（0.746） was found between dry weight per plant and plant height, whereas the least （0.349） was found between dry weight per plant and leaf width. Principal component analysis yielded three main traits: plant height, tillering, and leaf width. In principal component analysis, the contribution rates of the first three were 45.855%, 12.924%, and 11.594%, respectively, with a cumulative contribution rate of 70.373%. In principle component 1, the eigenvectors of the evaluated agronomic traits were all positive, and the load of dry weight per plant, plant height, stem diameter, number of leaves, and leaf length was higher. This means that high yield of Sorghum-Sudan hybrid grass was achieved when the plant was tall, the main stem was thick, the number of leaves was large, and the plant leaves were long. In path analysis, the direct effect on plant yield was plant height （0.444） > stem diameter （0.258） > number of leaves （0.201） > leaf width （0.136）. [Conclusion] Plant height and stem diameter are essential agronomic traits that have direct impacts on a single plant′s yield of Sorghum-Sudan hybrid grass.

Key words: Sorghum-Sudan hybrid grass, agronomic traits, yield per plant, path analysis