A field experiment was carried out for over two seasons (autumn and spring) as a split–split plot scheme based on a randomized complete block design with three replications. The main plots included two irrigation levels of the maximum available water depletion (maximum allowable depletion (MAD)) of 55 and 85% as non-stress and drought-stress environments, respectively, and the subplot accommodated two levels of nitrogen (0 and 62.5 kg N/ha, urea fertilizer); also, 20 barley genotypes were assigned to the sub-subplots. The biplot analysis of both sowing seasons showed that grain yield (GY) had a high positive correlation with total biomass (TB), whereas it had a high negative correlation with proline and total soluble carbohydrate as drought-tolerance-determinant characteristics. The genotypes which had the lowest and highest GY ranked significantly (P ≤ 0.01) different with changing the sowing season under each irrigation level, indicating a larger plant interaction and non-stability in response to the season change (about two-fold), as compared to the change in the irrigation conditions. It could also be concluded that barley genotypes might experience a higher decrease in GY and sensitivity to water deficit in the autumn sowing season, as compared to the spring planting season, which was also intensified by nitrogen application. However, the response to nitrogen application depends on the plant genotype.