BackgroundAs the most common primary bone cancer, osteosarcoma (OS) still lacks satisfactory therapeutic outcomes. Therefore, it is crucial to further evaluate OS at different risk levels and identify new intervention targets. Many evidences suggest the important role of angiogenesis in OS, but further exploration is needed.MethodsWe utilized public databases TARGET and GEO and employed bioinformatics algorithms such as LASSO, univariate and multivariate Cox regression analyses, and unsupervised consensus clustering to explore the role of angiogenesis-related genes (AGRGs) in OS. By calculating AGRG scores, we further analyzed OS molecular subtypes based on AGRGs. The correlation between AGRG scores and immune infiltration was subsequently examined. In vitro experiments, including WB, PCR, siRNA, migration, and invasion assays, were used to determine the value of the selected targets for OS.ResultsUltimately, we established an OS prognosis model based on five AGRGs (COL5A2, CXCL6, FSTL1, NRP1, and TNFRSF21) that can independently validate prognosis levels. In vitro experiments confirmed the aberrant expression of CXCL6 in OS and its potential role in migration and invasion.ConclusionOur study reveals the impact of angiogenesis on OS from a novel perspective and provides potential intervention targets.