In order to detect objects accurately,an image thresholding approach named two dimensions(2-D) maximum entropy was proposed to do infrared image segmentation.By using the 2-D histogram of image,the 2-D maximum entropy method not only considers the distribution of gray information,but also takes advantage of the spatial neighbor information.However,its great computation was often an obstacle in application.The threshold vector was obtained by using a new optimization algorithm,namely,the particle swarm optimization algorithm (PSO).The new way was proposed to realize the 2-D maximum entropy method instead of exhaustive search method.And it is 300~400 times faster than the traditional method.Through the example of segmenting the infrared image,the proposed method has been proved to be a fast method of segmenting infrared image.