Abstract: To address poor visual quality and difficulties in detecting dim-small targets in infrared image sequences due to temporal clutter, this paper proposes an algorithm that effectively suppresses temporal clutter and enhances image signal-to-noise ratio (SNR) without attenuating the energy of dim-small targets. This method cnstructs an adaptive filtering structure including temporal filtering, grayscale filtering, weight normalization, and clutter suppression estimation. It estimates the current frame by using historical multi-frame pixel grayscale values, calculates temporal filtering weights via a one-dimensional Gaussian function, and introduces a grayscale filtering module to eliminate motion blur in dynamic scenes, thereby smoothing the grayscale values at the same pixel position across consecutive frames. Experiments show that pixel values in non-target areas become smoother after suppression, while grayscale values remain largely unchanged when a target is present. The processed images show a reduction in mean temporal clutter from 1.56 to 0.83, achieving a clutter suppression factor of 2.27. The mean temporal SNR increases from 27.93 to 34.32, yielding an SNR gain of 1.20. The algorithm effectively reduces temporal clutter while successfully preserving the grayscale characteristics of dim-small targets, significantly improving image visual quality and target detection capability. It achieves a favorable balance among clutter suppression performance, computational efficiency, and target preservation, demonstrating high practical value and potential for widespread application.