The present investigation was undertaken to characterize the available TGMS lines and to explore the possibility of development of two-line rice hybrids at DRR, Rajendranagar, Hyderabad during Kharif 2002 and Rabi 2002-2003. Based on pollen sterility and spikelet fertility studies on sixty TGMS lines seven were found to be complete sterile. These seven TGMS lines were characterized with respect to sensitive stage, Critical Sterility Point (CSP) and Critical Fertility Point (CFP). The sensitive stage for fertility/sterility induction ranged from 15 days (DRR14S and DRR39S) to 24 days prior to heading (DRR1S and DRR 29S). The CSP was varied from 37.2oC (DRR25S) to 34.2oC (DRR39S) where as CFP was below 24oC for four TGMS lines and it was above 24oC in case of three lines. The seven TGMS lines showed a clear-cut tendency of high temperature (34oC / 26oC) sterile and low temperature fertile (26oC and 24oC/21oC) under growth chambers. Based on the morphological characterization, two lines DRR1S and DRR39S can be easily differentiated from other lines and based on auricle colour, ligule and leaf sheath colour these two lines can be separated from each other. Forty-two experimental two line hybrids were evaluated using Line x Tester design. Except panicle length all characters studied were governed by non-additive gene action. Two parental lines, DRR39S and R1 were rated as good general combiners for yield and yield related characters.Three hybrids, DRR 39S x R9, DRR 16S x R1 and DRR 25 S x R16 showed highly positive sca effects for yield. Three hybrids, DRR39S x R9, DRR 16S x R1 and DRR29S x R5 showed significant positive standard heterosis over KRH-2. Molecular characterization with 14 primer pairs revealed that among 14 lines, 38 alleles were found to be polymorphic and on average 2.85 alleles were amplified per primer pair. Polymorphic information content values varied from 0.565 to 0.948. Two primers, RM16 and RM25 generated unique alleles in two TGMS lines-DRR 23S and DRR6S respectively. The UPGMA analysis clearly grouped 14 lines into two major clusters (MAI and MA II) and more than 59% variation in the estimates of genetic similarity was revealed by PCA analysis.