Gboma is a dual-use traditional vegetable with a dense nutrient composition, a high tolerance to many biotic stresses, and a high market value. However, it remains neglected in modern breeding, with a paucity of knowledge of the genetic diversity of the genebank materials. To support breeding efforts, we analyzed the genomic variation in the bulk of accessions available in the university genebank (CalaviGen) and the World Vegetable Center genebank to select a representative panel to reduce redundancy, lower phenotyping costs, and provide breeders with a strategic subset of accessions for trait discovery and parent selection. For this purpose, we analyzed the genetic diversity and population structure of 290 gboma accessions, comprising 286 accessions from West Africa and four commercial varieties. Of the 1,490,934 SNPs identified, 18,628 high-quality SNPs were retained for genetic analyses. Cluster and structure analyses revealed two major clusters and two sub-populations, respectively, whose genetic structuring were independent of climatic zones and countries. All commercial varieties fell within cluster I and subpopulation I, which accounted for 18.28% and 16.9% of the whole collection, respectively. By integrating SNP data with 30 quantitative and 12 qualitative traits, we developed a representative diversity panel of 38 accessions that captured more than 99% of the allelic richness of the full collection. This study highlights the extent of genetic variation and exchange across West Africa, the limited overlap between commercial varieties and the broader germplasm, and the availability of a representative panel for research and breeding for economically and agronomically important traits in gboma.