Abstract: The aeural-fuzzy system has been applied to acquire the constitutive relationships of endogenetic particle distribution in FGM creating centrifugal casting at different mould temperatures, pouring temperatures and rotating speeds. During building up the neural-fuzzy model of the constitutive relationship for the alloy, mould temperature, pouring temperature and rotating speed were taken as the inputs while the relative thickness of the endogenetic particle distribution in FGM was taken as the output. Comparison of the predicted and experimental results shows that the neural-fuzzy system model used to predict the constitutive relationship of the endogenetic particle distribution in FGM creating centrifugal casting has good learning precision and good generalization. It is viable to forecast the endogenetic particle distribution in FGM creating centrifugal casting based on the neural-fuzzy system.