At first, two kinds of miniature LRB models are prepared, one model has normal lead damper, the other has abnormal one. And they are scanned by ultrasonic wave using ultrasonic transducers and AE sensors. The output signals obtained through the inside are different from each other. Infinite route is assumed as ultrasonic wave propagation line. But, in this examination, we found that the main route is some layers of steel. Then, the tomography construction is adopted to detect the condition of lead plug in one layer of steel plate. The full acoustic wave solution in an arbitrary two-dimensional object is computed, which is subjected to specified acoustic sources and material parameters. This algorithm is basically a finite difference method. As a result, it is successful to know about ultrasonic wave propagation in one layer of steel and consistent correlation with the output signal. The feature vectors are created by calculation of second power average of the amplitude, and these vectors are used as the data for pattern recognition on the basis of these information. The accuracy of pattern recognition is significantly enhanced in consistent correlation with the condition of lead plug. In this study, active damage detection method to evaluate the status of LRB was proposed. It is possible to evaluate internal states of LRB by using ultrasonic wave.