Water and nitrogen (N) are the most important limiting inputs of maize (Zea mays L.) production in semi-arid environment. Irrigation regimes, N application timing and rate significantly influence maize growth and yield. To study effects of the nutrients in semiarid environment, two experiments were conducted. The crop was sown on August 1, 2009 and August 2, 2010 at the Agronomic Farm of the University of Agriculture Faisalabad, Pakistan. The Cropping System Model (CSM) CERES-Maize was used to predict crop growth and yield under semi-arid environment. In Experiment-I, effects of times of N application were studied with five N levels (100, 150, 200, 250 and 300 kg ha-1). In Experiment-II, effects of three irrigation regimes (normal, water deficit at vegetative and water deficit at reproductive stage) were studied with the five N levels. The times and rates of N significantly affected grain yield and the yield components. Effect of N was linear, quadratic and cubic. Results revealed that application of N at the rate of 250 kg ha-1 in three splits (at 2 leaves, 16 leaves and silking stages) produced the highest grain yield (8.38 t ha-1). At these levels maximum values of leaf area index (LAI), photosynthesis and all others yield components were achieved. In contrary, crop growth rate and biological yield increased up to 300 kg N ha-1. However, the biological yield at 300 kg N ha-1 was statistically similar with 250 kg N ha-1. The crop was more sensitive to water deficit at reproductive stage than at vegetative stage. Water deficit at vegetative stage reduced grain yield 14% while of grain yield was decreased 22% when crop faced water deficit at reproductive stage. The CSM-CERES-Maize application predicted crop growth and yield well. So, the model can be used as a research tool in semi-arid zone of Pakistan. The model predicted phenology of crop well with less MPD and RMSE. The model predicted grain yield very closely to that observed; RMSE ranged from 656 to 1586 kg ha-1 among all treatments. In nutrient stress treatments, time course simulations of the model for LAI were satisfactory in both experiments. Prediction of the model was excellent in treatments which had adequate nutrients supply. The value of d-statistics ranged from 0.56 to 0.99 for LAI. The model predicted TDM very close to the observed values having low RMSE with 0.97 to 0.99 d-statistics. The application of 250 kg N ha-1 in three splits i.e., 1/3rd N at V2, 1/3rd N at V16, and 1/3rd N at R1 stages with supplemental irrigation of 530 mm ha-1 is the best management practice for semi-arid environment.