Currently, there is no generally-accepted defensible methodology that can be used to compare the energy performance of micro-cogeneration systems against reference conventional methods for providing residential thermal and electrical demands. In particular, the mathematical models of micro-cogeneration systems being used in this field have uncertainties associated with their predictions. However, these uncertainties have yet to be given serious attention by researchers in this field. This research makes a contribution by developing a more defensible methodology where these model uncertainties are considered. This methodology is demonstrated with a case study where the energy performance of a fuel-cell based micro-cogeneration system serving only domestic hot water demands is compared to the energy performance of reference conventional methods. A condensing tankless water heater and a central gas-fired combined-cycle plant are considered as these reference conventional methods for providing domestic hot water demands and electrical demands respectively. The simulation results demonstrated that if model uncertainties were ignored, it would have been possible to demonstrate that the considered micro-cogeneration system was more efficient than the reference conventional methods by serving only average consumption levels of domestic hot water. However, if model uncertainties were considered, the micro-cogeneration system could not reliably be shown to be more efficient than the reference conventional methods by serving the domestic hot water needs of a single house.