Next-generation accident-tolerant fuels for light-water reactors (LWRs) are being developed with high thermal conductivity additives. To reduce the high centerline temperatures that occur in present LWR fuel, complex fuels with improved heat dissipation capabilities are the goal. Depending on the manufacturing process, these fuels can use a variety of additives, and their microstructures vary. The thermal conductivity of UO2-SiC complex fuel is predicted using the finite element method (FEM) and machine learning (ML), as shown in this research. The method accurately predicts the corresponding complex fuel thermal conductivity and captures the characteristic value successfully. Novel anisotropic fuel pellets with higher thermal conductivity in a desired direction can be designed through the application of FEM and ML to material design. Data from published experiments with UO2-SiC complex fuel are used to validate the model and methodology. The model finally predicts the ideal characteristic parameters of the UO2-SiC based on the expected thermal conductivity.