Nondestructive detection for standing tree trunk is more difficult than that for other materials because of characteristics of radar wave imaging of standing tree trunk itself. According to the characteristics like complex internal structure and wide individual differences among different tree species, a method of absolute position characterization of internal defects was proposed. This amplitude ratio was used to estimate relative dielectric constant, and the Hilbert method was used for horizon tracking to realize the relative localization of defect interfaces. Then, the relative position was combined with the contour cloud data of standing tree trunk. The absolute coordinate of the internal structure of trunk was located and wood faultage image was reestablished at last. The internal structure analysis software was designed based on radar wave. For verifying the feasibility of the method, three experiments were performed: numerical simulations using the software GPRMax2D, based on finite-difference time-domain method, laboratory log measurements and real tree trunk measurements. The result showed that this method can locate the target point accurately. Its image resolution can show the internal defects such as holes and rots, and the results coincided with the actual structure of standing tree trunks during the restoration in the Summer Palace.