This paper presents an optimal management control strategy for power systems in industrial plants. A dedicated code has been developed to perform system analysis and simulation. The energy/mass balances existing between building and power plant has been depicted through a mathematical model based on vector equations, taking into account the behaviour of each system component. The main result is the definition of the power plant component set points satisfying the energy load under predefined optimization criterion (i.e. system efficiency, costs, pollutant emissions). Input data are the industrial plant loads, both electric and thermal, the technical characteristic of the installations, and the cost of electricity and fuel. As a general result we show that the optimal management of a power plant is as significant as the efficiency of its components for energy saving purposes. In particular, the correlation between the component set point profiles and the energy/cost/pollution savings is highlighted. Yearly simulations are performed on an existing energy system of an industrial plant varying the frequency of energy load dataset. The considered time steps are month, half a day, 4 h and I h. The results demonstrate that the whole power plant management leads to a global reduction of the cost and that the availability of more detailed energy load dataset leads to better operation cost estimation. As expected, considering a large time-step, the variation of energy load is not appreciable. The energy saving potential of this method is demonstrated allowing the best plant management solution under different energy loads. (C) 2008 Elsevier B.V. All rights reserved.