In any breeding program, it is necessary to screen and identify phenotypically stable genotypes that could perform uniformly under different environmental conditions. Such a breeding effort requires basic information of genotype × environment (G × E) interaction. Twenty genotypes including hybrids and aromatic rice were evaluated in 8 environments in two production systems viz; System of Rice Intensification (SRI) and normal cultivation environments during kharif season (May–October) 2009. The experiment was laid down in RBD with two replications in a plot of 1 m2. Pooled analysis for G × E interaction and stability revealed that the genotypes and environments were highly significant (p < 0.01) for all twelve characters studied. The G × E interaction was significant for six traits including all key components of SRI except tillers no. Both linear and non-linear components contributed towards G × E interaction. Stability parameters identified genotypes PR-114 and HKR-47 as stable for grain yield per plant and HKR-127, HKR-120, CSR-30, Pusa-1121 and IR-64 for test grain weight. Genotypes identified suitable for favourable environments were HKR-126, HSD-1, PAU-201and Govind while for unfavourable environment were HSD-1, HKRH-1094, HKR-48 and PAU-201 for different traits. IR 64 and Pusa 1121 registered 24.31 and 12.54% increase in yield in SRI over normal production system. These genotypes need to be tested in macro environments over space and time and could be utilized for direct cultivation as well as for improvements of other cultivars.