To reduce the effect of genotype × environment interaction and perform more accurate selection, performance and stability of genotypes should be considered simultaneously. In this study we used parametric and nonparametric methods to determine stable genotypes and the correlation between these methods was discussed. Ten lentil (Lens culinaris Medikus) genotypes in five regions of Iran (Gachsaran, Kermanshah, Shirvan, Gonbad and Ilam) were evaluated for yield stability within two years (2003-2004). Based on stability analysis by methods of Eberhart and Russel, Lin and Binns, Wricke Ecovalence and coefficient of determination of Pinthus introduced, genotypes of (9, 6), (5, 9), (8, 9) and (9, 8) were identified as the most stable ones, respectively. Results of analysis of Thennarasu and Nassar and Huhn non parametric statistics showed that genotypes of (9, 8), (9, 8, 1), (8, 9), (2, 1) and (1, 2) were identified as stable genotypes were stable genotypes, respectively. Finally by the correlation between parametric and nonparametric methods, genotype 9 (ILL 6199) was selected as a stable genotype. Results of this research showed that If the goal is to determine general compatibility, Shukla variance and Mean of Squares of Lin and Binns parametric statistics and NP(1) and NP(3) Tennarasu and Si(2) Nassar and Huhn nonparametric statistics are more proper than other statistics that studied in this research.