Abstract: Rice is one of the world's most important staple crops. Traditional rice genomic research has long relied on a single reference genome (e.g., Nipponbare), which suffers from "reference bias" and fails to capture the full breadth of genetic diversity in cultivated rice and its wild relatives. With the rapid advancement of long-read sequencing, graph-based pangenome assembly, and other technologies, rice pangenome research has entered an accelerated phase. This review summarizes the major advances, key research findings, current challenges, and future directions in rice pangenomics, including: (1) Innovations in high-quality genome sequencing and pangenome construction methods. (2) Novel insights into rice genetic diversity and structural variation from pangenome analyses, revealing domestication-related gene loss and transposon-driven phenotypic diversity. (3) The application of pangenomics in molecular breeding, such as structural variation marker development, elite allele mining, breeding design and AI-assisted breeding. (4) The challenges associated with large-scale data processing and functional gene validation, as well as future directions such as AI-driven breeding design. Overall, the rice pangenome provides new tools and paradigms for deciphering the genetic basis of complex traits and accelerating variety improvement, and offers important insights for pangenome research and molecular breeding in other crops.