Hybrid rice has been cultivated worldwide since Yuan LongPing initially proposed the idea of developing and applying hybrid rice in rice production 50 years ago. Hybrid rice has provided a strong guarantee for the food security in China. The application of hybrid rice was firstly based on cytoplasmic male sterility using a 3-line system, including the sterile line, the maintainer line, and the restorer line. Later, the discovery of some environment-sensitive genic male sterile lines whose fertility changed in different environments, indicated their strong application potential in hybrid rice production. Shi MingSong accidentally found a natural male sterile mutant Nongken 58S in the field when planting normal japonica variety Nongken 58. The pollen of Nongken 58S was sterile under long-day conditions, while was fertile under short-day conditions. Based on this special property, Nongken 58S could be developed as a photoperiod-sensitive genic male sterile line for hybrid rice: it would be a sterile line at longdays, meanwhile served as a maintainer line at shortdays. Since then, the theory of 2-line system has been improving. Furthermore, the discovery of some other thermosensitive genic male sterile lines such as 5460S, Annong S-1 and Hengnong S-1 enriched this system. They were sterile under high temperatures, but converted to fertility under low temperatures. The 2-line hybrid rice based on the photoperiod thermo-sensitive genic male sterile lines has played important roles in the hybrid rice production. From 1996 to 2013, the proportion of 2-line hybrid rice increased from 0.92% to 33.59%. Some early genetic analysis indicated that the fertility of most photoperiod thermo-sensitive genic male sterile lines was controlled by a recessive allele. And the genetic linkage map of these loci has been established. But only until recent years, with the completion of rice genome and the development of new sequencing technologies and rice transgenic techniques, several genes controlling photoperiod- sensitive genic male sterility or thermo-sensitive genic male sterility in rice have been cloned. Among them, pms3 was the first reported, the mutation of which caused Nongken 58S. pms3 encodes a long non-coding RNA LDMAR. A SNP of LDMAR altered its RNA secondary structure, resulting in increased siRNA-directed DNA methylation in the promoter, which reduced the expression of LDMAR in Nongken 58S under long-day conditions. Another group found that the SNP was located in a siRNA (termed as osa-smR5864m); loss-of-function of this siRNA may fail to suppress the expression of its target genes. Shortly after, tms5 , a thermosensitive genic male sterility gene in Annong S-1 was charactered to encode RNase ZS1 processing Ub L40 mRNAs. Overaccumulation of Ub L40 mRNAs resuting from a premature stop codon of RNase ZS1 gene created by a SNP in the tms5 mutant caused the male sterility under high temperatures. Elucidation of the molecular mechanism of these genes have important implications for breeding new types of excellent and stable two-line male sterile lines in the future.