To identify high-quality cotton varieties with superior agronomic performance， a three-year comparative trial was conducted using 24 early and early-mid maturing upland cotton varieties. Multiple statistical analysis， including correlation analysis， principal component analysis （PAC）， clustering， and gray correlation analysis， were employed to evaluate phenotypic traits. The results show that the coefficient of variation across the three years study ranged from 0.21% to 4.18%， with 2022 exhibiting the highest variability， particularly in the number of bolls per plant and single boll weight. Trait association analysis revealed 17 significant and 37 highly significant correlations， suggesting complex interactions between agronomic traits， yield components， and fiber quality parameters. Fiber quality traits showed stronger inter-trait correlations than other trait combinations. PCA revealed four principal components accounting for 77.98% of the variability. Based on the affiliation function， gray correlation model and AHP model， an integrated evaluation system was developed. This system identified superior cultivars， such as Jinfeng 6， J8031， and Xinnongda Cotton 1， based on their comprehensive performance metrics. The comprehensive evaluation system was validated within a panel of 283 upland cotton germplasm resources and 416 self-bred elite lines. Through this screening process， high-quality materials such as A191， A110， Y228， Y210， and Y297 were identified， along with materials with poorer overall performance， including A241， A39， Y110， Y366， and Y329. Practical field tests confirmed that the evaluation results were consistent with the actual performance in the field， demonstrating that the evaluation system has a solid foundation for further promotion.