Sixteen newly evolved upland cotton strains developed through conventional and mutation breeding techniques along with two commercial check varieties viz., CIM-496 and CRIS-134 were evaluated two years at five locations in province of Sindh, Pakistan. Information was recorded on seed cotton yield from all location and analyzed statistically. To assess the stability for yield between genotypes and stomata studies was performed. The results from the combine analysis of variance depicted that the mean square for seed-cotton yield for genotypes and genotype x environment interaction (G × E) were highly significant. Genotypes NIA-M-30 showed the maximum seed cotton yield (3.509 ton/ha) with low regression coefficient (b=0.893) and deviation from regression coefficient (S2d=0.015). Other four genotypes viz., NIA-80, NIA-83, NIA-Bt-2 and NIA-Perkh also produced higher yield (3.36, 3.32, 3.31and 3.25 ton/ha respectively) with regression coefficient (b=0.972, 0.918, 0.99 and 0.916 respectively) and deviation from regression coefficient (S2d=0.025, 0.003, 0.00 and 0.024 respectively), indicated wide adaptability to the range of environments. While the cotton genotypes NIA-HM-327, NIA-84 and NIA-HM48 showed high regression coefficient (b= 1.204, 1.331 and 1.291 respectively), which suggest their specific adaptation to favourable environments.