Soybean (Glycine max (L.) Merrill) production and productivity in Ethiopia remains low compared to the world average productivity potential due to lack of diversified soybean genotypes. As a result, genotypes that have been introduced from abroad need to be characterized to generate information for further breeding programs. Eighty one genotypes were tested in 9x9 simple lattice design to assess their variability and association of traits using quantitative, qualitative and quality traits at Pawe main station and Dibate substation during 2018/2019 cropping season. Based on analysis of variance, most of traits showed highly significant (p<0.01) differences at both locations except number of nodules per plant, number of pods per plant and number of seeds per pod. High GCV and PCV values were observed by number of seeds per pod and grain yield at Pawe and by grain yield at Dibate. High heritability coupled with high genetic advance as per cent of mean was recorded by number of seeds per pod (82.8% and 56.8%) and grain yield (89.9% and 45.2%) at Pawe and by days to flowering (97.2% and 23.7%) and plant height (96.6% and 33.7%) at Dibate. Number of pods per plant showed highly significant (0.35**) and significant (0.3*) genotypic correlation coefficient with yield at Pawe and Dibate, respectively. Highly significant and significant positive phenotypic correlation with grain yield was recorded by number of pods per plant (0.20*) and (0.24*) at Pawe and Dibate, respectively. Positive and high direct effect at genotypic level was observed by number of pods per plant (0.47) and number of seeds per pod (0.13) at Pawe and by number of pods (0.38) and number of seeds (0.20) on grain yield at Dibate. Highest phenotypic direct effect on grain yield was observed by number of pods (0.44 & 0.48) at Pawe and Dibate, respectively. About 0.83 and 0.88 residuals were observed from genotypic direct and indirect effects at Pawe and Dibate, respectively. Whereas, 0.84 and 0.90 residuals were observed from phenotypic direct and indirect effects at Pawe and Dibate, respectivel . The first four principal components accounted for 63% and 65 % of the total variation at Pawe and Dibate, respectively. Cluster analysis showed the existence of four divergent groups and the maximum inter cluster distance was observed between cluster I and cluster IV (D2=875.31)) at Pawe. At Dibate, the maximum inter cluster distance was observed between cluster II and cluster IV (D2=1227.68). However, one season experiment at two locations would not realize genotypes’ variability in response of environment, because quantitative traits are polygenic and profoundly influenced by the environment. Thus, further experiment on changed seasons and more locations is required. Keywords: Correlation, Genetic variability, Genotypes, Heritability, Soybean vi TABLE OF