In recent years, due to the emergence of military and civilian applications of the outbreak of growth, the unmanned aerial vehicles and light helicopters as the representative of the low-altitude slow small target has been rapid developed. Because such targets have the characteristics of “low flying altitude, slow flight speed and small RCS”, the target echo is often submerged in the strong ground/sea clutter. The traditional detection technology based on the frequency domain filtering is difficult to detect, which brings new serious challenge for the air defense early warning radar. To solve the problem, the multi path interference, atmospheric attenuation, system loss and other factors, the construction of the target echo model are considered; based on the strong clutter suppression by Kalmus filter, the distance and slow time image of the multi-frame radar echo data under long term observation is simulated by the slow time domain. After combining with the Radon transform, the target detection is completed through the threshold setting of the transformed parameter field. The experimental results show the effectiveness of the proposed method.