This paper is concerned with the evaluation of an adaptive predictive control strategy developed for a class of nonlinear delayed discrete time systems. These nonlinear models have the capability of describing plants whose dynamics depends upon the operating point. For this class of nonlinear systems, one considers a indirect adaptive predictive control scheme. The approach considered is evaluated through an application to a heat exchanger. Comparison with adaptive predictive controller using linear model is made to show the improvement of performances obtained by making use of the nonlinear scheme.