Grain yield stability measurements under varied environmental conditions suggests selection of suitable genotype. Safflower genotypes having different origins were tested for G x E effects on yield using univariate and multivariate statistics in six test environments (three varied climate conditions and combination of years) during 2016-17 and 2017-18. Genotypes were classified as stable or unstable based on multiple stability measures. Comparative yield results reveals different genotypic grain yield response in each environment. High yield (1618.4 kg ha-1) performance and responsiveness of PI-16308 genotype to favorable environment and lowest 1419 kg ha-1 was observed for PI-26748. In lieu of varied climatic condition Env4 showed 37% extra grain yield compared to Env3. The PCA results placed Env1, Env3 and Env5 in one group for producing similar grain yield and rest of the environments showed highest variation. Results. Of parametric (S2di) and non-parametric (S(3), S(6), NP(1), NP(2), NP(3) & NP(4)) measures indicates that PI-16308 and PI-16315 can be used to improve stability. However, Wricke's ecovalence (Wi2), Shukla's stability variance (σi2), θi, θ(i) , S(1) and S(2) could be considered for improving grain yield of PI-26748 genotype. All rights reserved, Elsevier.