Synthetic aperture radar (SAR) is valuable military tool that can generate high-resolution images in any illumination or weather condition. As SAR can be used for hostile purposes, it is important to develop jamming techniques to degrade the quality of SAR images. Jamming performance evaluation (JPE) metrics are used to evaluate the effectiveness of jamming systems. Existing JPE metrics are using the Euclidean distance or correlation coefficients, but they can vary significantly depending on the reference image. This makes it difficult to intuitively evaluate noise-jamming performance. In this study, we propose a new JPE metric based on the detection rate through SAR-automatic target recognition (ATR). The proposed metric uses the detection rate to directly infer the quality of the SAR image, which is the main objective of jamming. This allows for the direct evaluation of SAR image validity using the proposed metric.