Abstract: In radar echo signals from forest-covered scenes, issues such as low signal-to-noise ratio and fluctuations in signal amplitude and phase significantly increase the difficulty of target detection. To meet the application needs for detecting forest-covered targets with low-resolution radar at the signal level, a moving target detection method based on the fusion of full polarimetric radar echo signals using improved principal component analysis (PCA) is proposed in this paper. In this method, moving target signals are first extracted from the clutter background, and the fusion of full polarimetric radar echo signals is performed using the improved PCA. Detection is then carried out in both time and range dimensions, followed by re-detection through a non-coherent accumulation method, which effectively eliminates target overlap and false alarm interference. The target is ultimately detected, and its region of interest is also extracted. Finally, the method is experimentally validated using an autonomously developed L-band fully polarimetric radar system. Experimental results demonstrate that the proposed method significantly enhances the detection performance of the L-band fully polarimetric radar in forest-covered scenes, achieving excellent detection results for moving targets.