Drought is a basic reason of low grain yield in many field crops including maize (Zea mays L.). Identifying growth stages of any promising cultivar for irrigation scheduling under local climate and soil fertility can maximize yield. Potential soil moisture deficit (PSMD) approach (a difference between potential evapotranspiration and rainfall plus irrigation) describes the response of canopy growth to water shortage. To examine the productivity of different autumn-sown maize hybrids under irrigation scheduling at different growth stages and PSMD levels, a field study was conducted with split plot arrangement at the Agronomic Research Area, University of Agriculture, Faisalabad (Pakistan) during 2009 and 2010. The treatments consisted of two maize hybrids (Monsanto-919 and Pioneer-30Y87) and seven irrigation levels: I 1 = no irrigation (control), I 2 = 4 irrigations, I 3 = 6 irrigations, I 4 = 8 irrigations, I 5 = 10 irrigations (1 st irrigation starting at 6 leaf stage, followed by subsequent irrigations at an interval of 4 leaf stages in all irrigation treatments and addition of two reproductive stages in I 3 , I 4 and I 5 successively), I 6 = irrigation at 25 mm PSMD and I 7 = irrigation at 50 mm PSMD. The results showed that Pioneer-30Y87 produced more plant height but the two hybrids did not differ significantly in other growth and yield parameters. Among all irrigation treatments, 6 and 8 irrigation treatments were statistically at par with treatments of 10 irrigations and irrigation at 25 mm PSMD in radiation use efficiency, water use efficiency, 500-grain weight, grain yield and TDM production. So more number of irrigations is not a criterion for getting maximum yield in all maize hybrids. PSMD can be used as a useful criterion for scheduling irrigation under irrigated conditions of Pakistan. CSM-CERES-Maize model was used to simulate the growth, development and yield of maize under different irrigation regimes. The correct definition of soil drained upper (SLL) and lower limits (SDUL) played vital role in simulation by the model during dry season of 2009 and wet season of 2010. During model calibration, simulated total dry matter, maturity yield and maximum LAI were within 10 % of the observed values. During model evaluation, there was generally satisfactory agreement between all observed and simulated values of the two hybrids. Mean-Gini Dominance (MGD) analysis confirmed the observed information through 30 years climate data analysis that I 3 and I 4 were the best irrigation levels. So CSM- CERES-Maize can be used to simulate maize growth and yield at different irrigation regimes under semi-arid conditions of Pakistan.