A global supply chain is a complex network that must overcome several uncertainties. Supplier selection is an important attribute that is crucial to the company’s smooth supply and brand building. The article proposes two methods for optimizing supply chain finance decisions by diversifying supplier risk portfolios using graphical representations of payment performance relationships. These strategies incorporate constraints into convex portfolio optimization models using minimum spanning trees (MST) or triangulated maximally filtered graphs (TMPG). These additional constraints allow for the diversification of portfolios by selecting suppliers based on network centrality measures and asset neighborhood connections. A two-step stochastic optimization model is used to calculate the optimal risk portfolios for suppliers, utilizing payment indicators. The model then categorizes suppliers into four segments based on their portfolio risk-return tradeoff. Unlike traditional models, these methods prioritize peripheral suppliers and those not directly connected. The techniques were applied to a public dataset of supplier-payment practices, and the model could detect variability in payment behavior and generate optimized portfolios. These portfolios can then inform targeted financing strategies tailored to specific supplier segments rather than a generic approach. These methods improve supply chain finance decisions through graph-based supplier segmentation analysis.