A new method was presented for synthesis of infrared images of complicated terrain background and an image segmentation model was established based on the theory of Markov random field. By taking the effects of ambient and meteorological factors into account, the infrared features of different terrain scenes were simulated and calculated. The segmented visible pictures and the calculated values of the infrared features were synthesized by assigning the different gray scales to the pixels of the corresponding regions of the segmented pictures in order to obtain the infrared images of the complicated terrain scenes.