In order to advance the detection ratio of dim target in clutter background, the reason of low target detected radio from infrared image that contain both complex and flat background is analysed. And based on this analysis, an algorithm of dim target detection based on adaptive background adjusting is advanced. This algorithm uses Robinson guard filter to distill potential target point. And clutter image fuzzy membership function is defined to map the image to fuzzy plane. The background adjusting coefficient is computed by this fuzzy plane to modulate the potential target point. So the influence of the clutter background of the image is reduced. The experimental results have proved that this algorithm can obviously improve the probability of the point target detection, and it can detect dim targets within infrared image with signal-to-noise ratio higher than 1.