With this study, a new and more adaptable approach to the thermal design of the large heat sinks used in power electronics is proposed. This method, supported by the results from an extensive experimental program, recognizes that (1) the heat sink fins and the airflow adjacent to them form a simple cross-flow heat exchanger, and (2) conventional NTU-effectiveness methods can be adapted for use in the thermal analysis of the heat sink. This adaptation requires the development and evaluation of an equivalent heat capacity to describe the energy conducted along the fin. This method was initially used to evaluate the convective heat transfer coefficients between the fin and the cooling air. In this geometry, the developing airflow conditions make the prediction of representative values difficult. The correlation found to describe the test results was then used in an inverted analysis to predict and compare the experimental values for the heat sinks thermal resistance. The method is finally used in a design example where the fin spacing is optimized for a particular test duty. It is concluded that this new approach will make the design of large heat sinks more robust and reliable.