Because traditional background suppression methods don't take full advantage of direction information, small target and edges can't be distinguished effectively and some background may expose. A new method based on improved shearlet transform and image guide filter was proposed. Firstly, nonsubsampled wavelet transform and shear directional filter bank is adopted to decompose the infrared image into multi-scale and multi-direction to obtain high frequency and low frequency coefficients. Then the image guide filter is used to adjust the high frequency coefficients of each sub-bands to suppress the dim and small target signal and predict the background. Then the estimated background is obtained by the inverse transformation. At last, the suppression result is obtained by subtracting the estimated bcakground from the original image. Several group of experimental results demonstrate that the presented method can suppress the background and enhance the small target signal effectively, compared with several classical methods, such as Max-Median、TDLMS and Top-hat. The proposed method can effectively improve the detection probability of infrared search and tracking system for infrared dim and small target.