Organic Rankine Cycles (ORCs) are particularly suitable for recovering energy from low-grade heat sources. This paper describes the behavior of a small-scale ORC used to recover energy from a variable flow rate and temperature waste heat source. A traditional static model is unable to predict transient behavior in a cycle with a varying thermal source, whereas this capability is essential for simulating an appropriate cycle control strategy during part-load operation and start and stop procedures. A dynamic model of the ORC is therefore proposed focusing specifically on the time-varying performance of the heat exchangers, the dynamics of the other components being of minor importance. Three different control strategies are proposed and compared. The simulation results show that a model predictive control strategy based on the steady-state optimization of the cycle under various conditions is the one showing the best results. (c) 2011 Elsevier Ltd All rights reserved.