This paper reports on an empirical investigation into the ability of Artificial Neural Networks (ANN) in approximating group preference. Given the existence of a group utility function in a multi-criteria decision context, an ANN, as a universal function approximator, would be able to recognize group decision patterns, approximate the underlying functional relationship, and generalize over new cases. The merit of ANN over the traditional utility theory approach is in its ability to learn group preference without imposing strong assumptions on the functional form and behavior of decision patterns.