A novel technique of design of experiments applied to numerical simulations is proposed in this paper as a methodology for the sizing and design of thermal storage equipment integrated in any specific application. The technique is carried out through the response surfaces in order to limit the number of simulation runs required to achieve an appropriate solution. Thus, there are significant savings on the time spent on the design as well as a potential cost saving on the experimentation if similarity relationships between the prototype and the model are met. The technique is applied here to a previously developed and validated numerical model that simulates the thermal behavior of a phase change material-air heat exchanger. The incorporation of the thermal energy storage unit is analyzed in the case of a solar cooling application, improving the system coefficient of performance. The economic viability is mainly conditioned by the price of the macroencapsulated phase change material.