In the efforts to reduce both CO 2 emissions and non-CO 2 climate impacts in global aviation, encouraging the take-up of Sustainable Aviation Fuels (SAF) is seen as a key element. Although there are currently seven ASTM-approved SAF production pathways, availability of the required feedstock and increased costs compared to conventional jet fuels pose major challenges for a fast market uptake. It is therefore advisable to use the currently available capacities in a smart way, i.e. taking into consideration the maximum potential for mitigating climate impact for a given scenario as well as a cost-optimal and economic efficient allocation in the operational context. For such complex tasks, all information regarding the system under consideration should be made available over the entire lifecycle. For this purpose, a Digital Twin concept for SAF is proposed and demonstrated, linking together information from models and data to form a virtual representation of the fuel and mirror the fuel during its lifecycle. Special emphasis is placed on the representation of uncertainties of the involved model predictions and the traceability of the provenance of all information within the Digital Twin. The methodology is tested and demonstrated for a decision making problem typical for the utilization of SAF under the current regulatory constraints, namely the blending of conventional fuel with SAF to a drop-in SAF blend. The fact that the physical properties of the conventional jet fuel might vary within its specification requires a careful selection of the blend ratio and the candidate SAF to achieve an optimal blend given associated cost functions and constraints. It is identified that due to the defined constraints and the cost function, not always the SAF with the highest possible blend ratio is selected as the best choice. It is shown that the inclusion of uncertainties narrows the range of possible blend ratios and hence affects the decision process and the optimal decision. However, this approach reflects a risk-informed decision since uncertainties in the model prediction are considered in the decision process.