Global agricultural systems face mounting threats to food security from climate change, population growth, and land degradation, with current productivity gains insufficient to meet the 9.7 billion demand projected for 2050. Seeds, as both the carriers of genetic information and the starting point of agricultural production, directly determine crop yield, resilience, and quality. Advancing seed innovation is therefore central to achieving sustainable production gains. This review explores the evolution of seed science from agrarian civilization to the era of intelligent seed design, summarizes recent advances of AI-based methodological innovations and applications in seed science. It introduces the emerging paradigm of AI-driven seed design, summarizes core scientific questions and key technologies, and outlines technological pathways. Furthermore, it analyzes the challenges and future directions in this field. By integrating the latest research and technological developments, this review aims to establish an "AI for Science" paradigm in future-oriented seed science that can meet the increasing global demand for sustainable and high-quality seed resources.
ZHANG et al. (Sun,) studied this question.