Batch distillation processes are widely used in the chemical industry. In this work, we consider the optimization of such processes by simulated annealing. Although this method is stochastically in nature, it has two well-known advantages: it can be readily connected to highly sophisticated simulation codes and it converges towards a global optimum. According to the characteristics of batch distillation operation we propose to use a two-step computation approach. A feasible strategy (admissible control) will be searched for in the first step and it will be optimized in the second step. The approach has been applied to a model of a pilot batch distillation plant. These results show the potential of the method for developing optimal operation strategies for batch chemical processes;