To study the stability of seed yield of winter oilseed rape genotypes in target areas and to compare the types of stability parameters, experiments were conducted in randomized complete block design with 15 genotypes and a control cultivar (total of 16 genotypes) with three replications during the two growing seasons of 2015-2017 in four cold and semi cold locations (Kabutarabad-Isfahan, Torogh-Mashhad, Islamabad-e-Gharb, Karaj), Iran. Results of analysis of variance showed a significant difference between the oilseed rape genotypes in some locations. After confirming the homogeneity of variance of experimental errors in different environments, combined analysis of variance was performed with the assumption of the fix effect of genotypes and random effect of location and year. Results showed that the environment, genotype and genotype×environment effects were significant at 1% probability level and genotypes had different performance in different environments. The greatest contribution of the variations was related to the environmental effect (32. 83%). The statistical methods consisted of; environmental variation coefficient, stability variance, ecovalence, environmental variance, mean square deviations from regression line, coefficient of determination, regression coefficient Tai regression (Alpha and Lambda) and simultaneous selection index were used to determine the stable and high yielding genotypes. Although there was a possibility of selection for performance and stability by a functional criterion, but due to the existence of low and variable Spearman correlations between different types of indices, except for the simultaneous selection index and grain yield, a combination of different methods of stability was used and G4 (ES Alonso), G7 (Es Kamilo) and G16 (Okapi) genotypes were identified as stable and high yielding genotypes. It concluded that the selected genotypes could be used for cultivation in cold and mild cold regions of the country.