The scene model of infrared small target image was established, and the characteristics of image pretreatment were analyzed under this condition. Based on the characteristics, the pretreatment technologies were classified. In order to evaluate results of pretreatment, two definitions of signal-to-noise ratio and background suppression factor were given. The experimental results indicate that different pretreatment methods can be applied to different images. In order to get good pretreatment effects, the appropriate pretreatment method should be selected according to the image characteristics.