Most of the thermoeconomic accounting and optimization methods for energy systems are based upon a definition of the productive purpose for each component. On the basis of this definition, a productive structure of the system can be defined in which the interactions among the components are described by their fuel product. The aim of this work is to calculate marginal and average unit costs of the exergy flows starting fi-om their definitions by a direct inspection of the productive structure. As a main result, it is noticed that the only differences between marginal and average unit cost equations are located in the capital cost terms of input-output cost balance equations of the components.