Regarding the infrared target detection under the complex dynamic scene,a method is proposed for infrared moving target detection based on cross-entropy transition region extraction.Firstly,infrared image is processed with method of frame difference and background difference.Then binary image is obtained by segmentation algorithm based on cross-entropy transition region extraction.Finally,the complete infrared target is detected by the morphology filter.The experimental result indicates that this method is quite fit for infrared target detection and satisfies infrared moving target detection.