In recent years, my research has primarily focused on farmland habitat simulation. I have developed a dynamic crop yield prediction approach by integrating historical meteorologically similar year selection with crop models. I established a field irrigation decision-making algorithm based on dynamically predicted yields, and proposed a spatial upscaling framework for crop model cultivar parameters through stochastic sampling from prior distributions of cultivar-specific parameters. In addition, I evaluated the performance of empirical parameters of the Angstrom–Prescott (A–P) model at different temporal scales for incoming shortwave radiation estimation using machine learning methods, and generated a high-spatial-resolution dataset of A–P coefficients covering mainland China.