This paper investigates the problem of cross-range scaling of inverse synthetic aperture radar. It is very important to target classification and recognition via ISAR. The cross-range scale of ISAR depends on both radar wavelength and rotating angle of target relative to radar-line-of-sight (RLOS) during the total coherent processing time. The former is known while the latter is difficult to determine especially in the case of ISAR. In this paper, a new approach, which is based on the principle of tomographic imaging and the property of the coherent processing of echo data, is developed to estimate the rotating angle of target relative to RLOS during the total coherent processing time. As such,it is able to carry out the cross-range scaling of ISAR. The proposed approach is used to process the computer simulated data and the real data of model B-52 collected in a microwave anechoic chamber. The processing results show that the proposed approach is correct and effective. >