Cleaner energy production system such as direct alcohol fuel cells (DAFCs) are considered as an alternative source for generating cleaner energy. Studies based on design of catalysts, electrodes design, proton exchange membrane, and flow field were conducted for improving its performance characteristic such as power density. However, the less focus was paid on determining the operating conditions considering the uncertainties that will result in an increase of power density of DAFCs. Therefore, the present work proposes a novel comprehensive procedure involving experimental study and evolutionary approach of genetic programming (GP) in formulation of robust power density models for DAFCs. Two uncertainties such as the selection of objective function and variations in measurement of operating conditions are incorporated in framework of GP. The power density models incorporate the formulation of new objective function in GP that will result in higher accuracy of the models. Experiments performed on DAFCs validate performance of the models. Simulation profiler is then generated for models to verify its robustness in uncertain operating conditions. The inferences on relationships between power density and operating conditions for DAFCs are made by surface analysis of the models.