The work described in this thesis develops a strategy for planning domestic energy supply systems for small, rural communities. Two methods, for system optimisation are proposed: economic optimisation and thermodynamic optimisation. Both methods are applied to the small Scottish island of Eigg, as an example of a small community. It is shown that, for the range of energy equipment studied, the two methods are generally complementary. Economic theory and choice of parameters are discussed, as is the theoretical background to thermodynamic modelling. The net present value method is used to determine unit energy costs, which are used in the economic model as the parameter by which different energy sources are chosen. For thermodynamic assessment three parameters are used: exergy, second law efficiency and energy quality. Energy quality against quantity diagrams (on which areas represent exergy) are shown to be a useful way of visualising thermodynamic system performance. This work also develops methods which can be used to determine model inputs. These include: an energy census of the island of Eigg to determine the current energy system, development of a hydrological prediction model for determining hydro power potential, use of computer models to determine house space heating demands, and development of methods for predicting meteorological parameters from long term data. The application of the economic planning model to Eigg allows a much improved energy supply system for the island to be proposed. This makes energy more available, improves comfort (with regard to energy usage), and gives 30% higher energy usage for a lower unit cost than the islanders currently pay. The thermodynamic optimisation process improves the thermodynamic efficiency of the energy system on Eigg by more than 100% and shows which areas of energy technology should be investigated to harmonise economic and thermodynamic optimisation.