Determining management zone (MZ)-specific optimal N rate is a challenge in precision crop management. The objectives of this study were to evaluate the potential of applying crop growth model to simulate corn yield at varying N levels in different management zones and estimate optimal N rates based on long-term weather conditions. Corn (Zea mays L.) yield data from 1995, 1997 and 1999 were used to calibrate a modified version of CERES-Maize (version 3.5) model to simulate historical spatial yield distributions in a commercial field previously divided into four MZs in eastern Illinois. The model performance in simulating corn yield for two hybrids (33G26 and 33J24) at five N levels (0, 112, 168, 224, and 336 kg ha-1) in two independent years (2001 and 2003) were evaluated. Economically optimum N rates (EONRs) were estimated based on 15-years of simulated yields (1989-2003). The model explained approximately 59% of the spatial and temporal variability in corn yield during calibration and the model performance was improved with the two years of N experimental data, explaining approximately 92.9% of the variability in corn yield across hybrids, years and management zones. Model estimated EONR for two hybrids (33G26 and 33J24) in different management zones varied from 70 to 230 kg ha-1. Applying the long-term average hybrid- and management zone-specific EONR across years did not consistently improve the 15-year average economic returns over a uniform N application. However, applying the hybrid-and MZ-specific EONR that maximized the 15-year average net returns across years improved economic return by an average of 22.30 and 13.70 $ ha-1 for 33G26 and 33J24, respectively. More studies are needed to develop, test and apply crop growth model-based precision N management strategies.