Abstract: In this paper, the clutter suppression algorithm of external emitter radar is studied based on KL transform in wavelet domain. In radar records, the traditional KL transform filter is uncorrelated in the whole time axis and lacks analysis of some details in a certain period of time. When recording radar data, an improved algorithm based on KL transform, wavelet domain name “KL conversion method”, will be proposed. The improved algorithm can completely filter random noise and horizontal wave, highlighting the effective wave. Through theoretical data and measured data processing, KL transform filtering algorithm can effectively suppress direct wave and clutter, and the suppression performance is better than LMS algorithm. Compared with LMS algorithm, KL transform filtering algorithm has smaller misalignment error and faster convergence speed. In the calculation of complexity, KL algorithm is much less than LMS algorithm.