Abstract: This study aimed to enhance the performance of the WOFOST crop model in simulating spring maize growth and yield under varying water conditions, with particular emphasis on improving the representation of drought-induced water stress. The research focused on addressing the model's limitations in capturing soil moisture effects and developing a dynamic, adaptive framework for yield forecasting based on data assimilation and parameter optimization. Multi-year, multi-sowing-date field experiments were conducted at the Jinzhou Agrometeorological Experimental Station in Northeast China, covering three distinct moisture scenarios: normal water years (2011-2013), typical drought years (2014, 2015, and 2018), and mild drought years (2019, 2023, and 2024). This comprehensive dataset provides a robust foundation for systematically evaluating the model's performance across different water regimes. An Ensemble Kalman Filter (EnKF) data assimilation approach was implemented using Leaf Area Index (LAI) and Soil Moisture (SM) observations obtained from field experiments. Key photosynthetic parameters, including the maximum leaf CO₂ assimilation rate (AMAXTB), were optimized through iterative calibration against observed biomass and yield data. Additionally, the soil water stress response function (RFWS) was substantially modified to better reflect the nonlinear effects of soil moisture deficits on photosynthetic processes and biomass accumulation. The modification addressed the oversensitivity of the original function to slight moisture reductions below critical thresholds, which had previously led to excessive yield underestimation under drought conditions. Multiple simulation schemes combining different assimilation and optimization strategies were systematically designed and evaluated. These included: S1 (original model), S2 (LAI assimilation only), S3 (LAI assimilation with parameter optimization), S4 (SM assimilation only), S5 (SM assimilation with parameter optimization), and S9 (SM assimilation with both parameter optimization and RFWS modification). Each scheme was rigorously evaluated by comparing simulated above-ground biomass and final yield with comprehensive field observations under both normal and drought conditions. Model performance assessment revealed substantial improvements through the integrated approach. In normal-precipitation years (2011-2013), LAI assimilation combined with parameter optimization (Scheme S3) achieved the highest accuracy, reducing mean absolute relative errors (MARE) of yield and above-ground biomass from 38.3% and 42.4% in the original scheme to 11.5% and 14.2%, respectively. This improvement demonstrates the particular effectiveness of LAI assimilation under adequate moisture conditions where leaf development primarily drives yield formation. In drought years (2014-2015), SM assimilation combined with parameter optimization and modified RFWS function (Scheme S9) performed best, lowering yield and biomass MARE from 54.1% and 32.7% to 34.1% and 18.0%, respectively. The 2014 case study was particularly insightful, where despite severe meteorological drought, actual yield exceeded expectations due to carry-over soil moisture from the previous wet year (2013). Scheme S9 captured this anomaly more effectively than other schemes, though some yield underestimation persisted, indicating remaining challenges in extreme drought simulation. Rolling forecasts implemented throughout the growing season indicated that approximately 60 days after the jointing stage represented the optimal prediction node, corresponding to the critical period of water demand for maize and the phase when yield components become largely determined. A differentiated assimilation strategy was developed based on a cumulative precipitation threshold of 149 mm during this period: Scheme S3 (LAI assimilation with photosynthetic parameter optimization) was applied when accumulated precipitation exceeded 149 mm, while Scheme S9 (SM assimilation with photosynthetic parameter optimization and modified water stress function) was implemented when accumulated precipitation fell below or equal to this threshold. Validation with 2015, 2018, 2019, and 2024 sowing dates showed that the differences between simulated and observed yields were -649 kg/hm² in 2018 and 58 kg/hm² in 2019, demonstrating the model's capability in capturing yield trends across varying water conditions. The combined framework of data assimilation, parameter optimization, and water stress function modification substantially enhanced WOFOST’s ability to simulate spring maize growth and yield across diverse water conditions. This framework provides a practical tool for yield prediction, drought impact assessment, and decision support under varying water conditions. Future work could focus on refining drought response mechanisms, incorporating multi-factor parameterization including temperature and radiation effects, and extending the framework to other crops and ecological regions to improve model stability and applicability.