Abstract An empirical model for predicting the thermal capacity of a coil operating in dry mode is investigated. It is based on the well-known e-NTU method. Thermal resistances are estimated by empirical correlations calibrated with manufacturer data by using two operating cases. Large number of operating cases by ranging the inlet states, the flow rates of both fluids, the altitude and the propylene-glycol concentration is considered in tests. The validity of several approaches of the empirical model is investigated for heating and dry-cooling operating cases. When considering pure water as heat transfer fluid, acceptable results were achieved without any correction of properties. Similar tests were conducted considering propylene-glycol aqueous solution as heat transfer fluid. The error indicators are significantly reduced when the variation of the fluid properties is taken into account. According to the large number of tests conducted, the approach considering the correction of properties of both fluids is a promising procedure for the simulation of the coil, operating either as a heater or as a dry cooler, in TRNSYS, EnergyPlus or other dynamic simulation tools. The validity of the empirical component model is also investigated by using available experimental data published in the literature using aqueous solution of ethylene glycol as heat transfer fluid. Although the number of experiments used is not large and the accuracy of some measured cases is not as good as desirable, the tested approaches with and without correction of properties provide acceptable results.