This paper introduces the use of a genetic algorithm as a neural network training paradigm. A case study of an intelligent gas/odor sensor is presented to demonstrate the method’s effectiveness.