Over the last several decades of space flight, spacecraft thermal system modeling software has significantly advanced, but the model validation process, in general, has changed very little. Although most thermal systems flown are successful, there is some evidence of model inaccuracy and thermal system overdesign due to the conservatism of the current (i.e., conventional) validation process. This work proposes a Bayesian-based Model Validation (BMV) methodology as a tailored framework that combines the state of the art model validation methods within the fields of Uncertainty Quantification (UQ) and Design of Experiments (DOE) to improve the thermal model validation process. BMV is implemented in a passive spacecraft radiator sample. BMV is shown to be a rigorous, systematic validation methodology that can identify and reduce important model uncertainties in a spacecraft thermal system. BMV increases knowledge of the system earlier in the project lifecycle when important design decisions are made by focusing research and testing efforts on critical system sensitivities.