A comprehensive and detailed characterization of genetic resources is crucial for both harnessing specific traits and protecting unique germplasm. The research was conducted at the plant breeding research field of the National Genebank, Nepal in 2021. This study aimed to evaluate the variability among soybean landraces from various regions of the country by analyzing agro-morphological traits. 77 soybean landraces, sourced from 24 districts across the nation were characterized by following augmented block design. 11 qualitative and twelve quantitative characters were recorded following the descriptor developed by Bioversity International. Descriptive analysis was done by using MS Excel 2019 and multivariate analysis of quantitative characters was done with OriginPro 2024 software and R Studio 3.5 version. Among the studied quantitative characters, yield (112.71%), 1000 grain wt. (38.45%) and plant height (27.28%) showed high variability among the landraces. A dendrogram was generated for the 77 soybean landraces utilizing the UPGMA clustering approach, which incorporates average linkage and Euclidean distance, to analyze 12 quantitative traits. The cluster segmentation organized the landraces into two distinct categories with a distance of 111.42. The soybean landraces studied showed considerable variability in various qualitative and quantitative traits. This observed diversity offers valuable potential for both breeding programs and direct applications. These promising landraces need further screening through multi-environment trials and the application of various biotechnological tools.