Field experiments were conducted in 1999 and 2000 to evaluate and adapt the Herbicide Application Decision Support System (HADSS) program for Texas Southern High Plains cotton production. Weed management systems (in glyphosate-resistant, bromoxynil-resistant, and nontransgenic cotton varieties) included trifluralin preplant incorporated (PPI) followed by (fb) HADSS postemergence-topical (POST) recommendations (PPI fb POST HADSS), HADSS recommendations alone (POST HADSS), and Texas Agricultural Experiment Station (TAES) recommendations for the Texas Southern High Plains. In both years, effective season-long weed control was achieved with all weed management systems in the glyphosate-resistant variety, but only the PPI fb POST HADSS and TAES weed management systems controlled Palmer amaranth and devil's-claw in the bromoxynil-resistant and nontransgenic varieties, compared with POST HADSS alone. No differences in cotton lint yield or net returns over weed control costs were observed with weed management systems across variety in 1999; however, in general, the glyphosate-resistant and nontransgenic varieties produced higher yields and net returns than the bromoxynil-resistant variety. In 2000, plots from the TAES weed management system produced higher lint yields than the plots of PPI fb POST HADSS recommendations in the glyphosate- and bromoxynil-resistant varieties, but plots of all management systems yielded similarly in the nontransgenic variety. In 2000, plots from the TAES system produced the highest net returns in the glyphosate- and bromoxynil-resistant varieties. In the nontransgenic variety, the PPI fb POST HADSS and TAES weed management systems produced higher net returns over weed control costs than the POST HADSS system.