Electrically stimulated coral polyps form a traveling wave of contraction from the site of the stimulation. A model of a coral nerve network was optimized to match one of the experimentally observed behaviors, a constant velocity of spread of excitation. We applied genetic algorithms to increasingly more complex models of a coral nerve net. In the first stage, individual neurons were optimized to respond with a spike to multiple, but not single pulses of activation. In a second stage, we used these neurons as the starting point for optimization of a 2-dimensional nerve net. This strategy yielded a network with parameters that reproduced the experimentally observed spread of excitation.