Target damage assessment is an important basis for commanders to make scientific decisions and command correctly. An image change detection algorithm based on Gaussian mixture model and EM algorithm is proposed. The damage of the target is evaluated by detecting the image change rate before and after the attack of the target, and the difference fusion image is constructed by considering the gray and texture characteristics of the image at the same time, and the Gaussian mixture model is constructed for the changed pixels and the non-changed pixels of the difference fusion image, and the model parameters are calculated by the EM algorithm; Bayesian decision with minimum error rate is used to detect the change of the difference fusion image. The experimental results show that the method is reasonable and reliable, and the target damage assessment is practical.