Cosmetic corrosion of painted automotive substrates is a complex phenomenon being a function of number of environmental variables and material properties. To address the need for reliable accelerated corrosion tests, a high performance corrosion chamber was built by VOLVO car corporation, Gothenberg, Sweden. Using a statistically designed program of experiments, excellent correlation between outdoor and laboratory simulations have been established using the VOLVO technique. Traditional methods for corrosion data analysis has been based on the use of well known statistical methods. In this paper, we have introduced Artificial Neural Networks (ANN) to study and establish complex relations between scribe creep data and the variables that govern cosmetic corrosion performance. Application of the ANN methodology as a predictive tool has been discussed.