The detection and tracking process of radar targets are often interfered by ground clutters, sea clutters and whether clutters. After the traditional Moving Target Indication(MTI) filtering and Constant False Alarm Rate(CFAR) detection, the residual clutters have almost the same two-dimensional distribution(range-time) as the target, and it is difficult to distinguish them. To solve this problem, a plots filtering method for radar targets based on clutter feature evaluation is proposed. Synthetically considering the difference of features in amplitude fluctuation, distance/pitch/orientation correlation and phase shifts, comprehensive feature factors are evaluated to distinguish radar targets from clutters. Finally, the proposed method is verified by using the real radar measured data, which indicates that the effective suppression rate of clutters is significantly higher than that of traditional methods.