A new method of classifying thermal comfort of young pigs was investigated that may be used to interactively control the micro-environment. The method uses the spectral features of swine postural images as inputs to a neural network which then classifies the corresponding thermal environment as cold, comfortable, or too warm for the pigs. The results show that the method has great potential as an interactive control tool to improve swine well-being and production efficiency. Further research and implementation of the algorithms for development of the innovative, behavior-based swine environmental controller are warranted.