We develop a pricing model for electric power service that accounts for customers＇ reliability preferences as well as the randomness of outages in the power-supply system. The model combines features of both reliability pricing and real-time pricing. A production cost simulation submodel is used to estimate expected total and marginal production costs on an hourly basis, as well as system reliability measured by the loss-of-load probability. These estimates are inputs to a welfare-maximization model with revenue constraints differentiated by customer class. The methodology is illustrated by a numerical application with typical power system and market data. The results show that the proposed pricing approach is significantly superior to spot and Ramsey pricing in terms of economic efficiency, energy conservation, and generation-capacity requirements.