Due to the problems to infrared image segmentation using Markov random fields,a method for infrared image segmentation based on triplet Markov fields using mixture gauss model was proposed.The assumpation of conditional distributions independences were overcome by introducing an auxiliary random field in Markov random field and considering the Markovianity of all the random processes in the triplet Markov fields.And the objects areas and background areas were distinguished by this auxiliary process.In there the Gaussian mixture model was used to be the triplet Markov fields probability model.The experimental results show the infrared image can be segmented well by the triplet Markov fields for complex background.And the ideal segmentation results were obtained.