The design of experiments is a highly effective offline quality improvement method to optimize the existing and new processes or products. In the literature, standard experimental situations have been paid a lot of attention. In a number of non-standard experimental situations, special experimental design techniques should be considered in order to conduct an experiment for design factors. Indeed, an I-optimal design, a computer-generated special experimental design, is a good choice to predict the mean and variance responses under non-standard experimental design situations. In this research work, an I-optimal design is selected to generate experimental design points for a non-standard experimental situation. Then, an I-optimal design-based dual response optimization model is proposed in order to obtain an optimum operating condition for design factors while minimizing the process variance as small as possible. Comparison studies are also conducted. Finally, a numerical example is conducted in order to illustrate the effectiveness of the proposed optimization model.