Additive main effect and multiplicative interaction (AMMI) and best linear unbiased prediction (BLUP) are two methods for analyzing multi-environment trials (MET). In this study, seven selected rice lines were evaluated along with two check varieties based on randomized complete block design in Tonekabon, Amol and Sari (Iran) in three growing seasons of 2011-14. To quantify the genotypic stability, the best linear unbiased predictions of the genotype by environment interactions (GEI) were estimated, and singular value decomposition (SVD), which is the basis of AMMI analysis, was performed on the resulting matrix. The likelihood ratio test (LRT) showed that the effect of GEI was significant on grain yield, number of tillers, thousand grains weight and panicle length. Therefore, due to the significant interaction of genotype by environment, BLUP analysis can be performed on this data. The biplot of first principal component (PC1) of the environment versus nominal yield showed that genotypes 7 ([IR 67015-22-6-2-(A37632) × (Amol3 × Ramzanalitarom)]39), 6 (IR67015-22-6-2-(A37632) × (Amol3 × Ramzanalitarom)]126) and 2 ([IR64669-153-2-3 - (A8948) × (4Surinam Deylamani)]2), due to the lowest scores of the PC1, had a small share in the GEI and had more grain yield stability. The biplot of grain yield versus WAASB, placed genotypes in four regions, so that genotypes in the fourth region, including genotypes 6, 7, 8 (Line 843, check variety), and 9 (Shirodi, check variety), were due to large value of response variable (high grain yield) and high stability (low values of WAASB) were very productive and had extensive stability. Identification of genotypes with weighted average of WAASB and response variable (WAASBY) criteria showed that genotypes 6 and 7 were high yields and stable. Based on the multi-trait stability index (MTSI), G6 was also selected as the best genotype in terms of grain yield, evaluated traits and stability of each trait. Totally, genotype 6 was stable and superior based on the results of all methods.