The fracture process zone (FPZ) significantly influences the damage tolerance of adhesively bonded composite joints, governing crack-growth mechanisms and migration. Existing fracture characterization approaches generally evaluate pure-mode behavior independently and extend these results to mixed-mode conditions using a power-criterion, such as the Benzeggagh-Kenane (B–K) criterion. This process assumes that FPZ-dependent mode-mix behavior from a standard mixed-mode test is transferable to another complex loading condition. This assumption remains unchecked for toughened adhesive joints, where FPZ morphology varies with loading conditions. This study addresses this gap through experimental and numerical investigation using digital image correlation (DIC) and cohesive zone modeling (CZM). The pure mode I test displayed localized FPZ ahead of the crack tip, influenced by carrier bridging. Two different pure Mode II tests demonstrated that the apparent crack length method accurately accounts for the large FPZ ahead of the crack tip. The mixed-mode bending (MMB) test linked pure modes through the B–K criterion. The Crack-Lap Shear (CLS) specimens exhibited evolving FPZ and mode II-dominated fracture. The fracture toughness predicted by the B–K criterion deviated from the CLS tests as the loading became more mode II dominant. It was observed that the FPZ morphology during the CLS test differed significantly from that observed during the MMB test, through DIC and CZM. These results highlight that differences in FPZ affect the mixed-mode fracture toughness and demonstrate the limitations of applying a single empirical power-criterion. It underscores for FPZ-sensitive approaches to accurately predict the fracture resistance of toughened adhesive joints under evolving mixed-mode conditions.