Nowadays radar plays an important role in both militaries and non militaries applications. One of the most important applications of radar is imaging. Rotational motion is essential for imaging. This motion can be provided by target or radar and sometimes both of them. So imaging radars are divided into two categories: SAR (Synthetic Aperture Radar) and ISAR (Inverse Synthetic Aperture Radar). In SAR imaging, target is stationary and radar is moving and generates image from the surface of the earth, other planets and so on. But in ISAR imaging radar is moving and target is stationary. Fourier transform is a conventional tool for ISAR processing. When the target is not maneuvering, this way leads to a focused image. But when target is maneuvering, unwanted phase error is produced and the FFT image is blurred. So, to provide a focused image, compensation algorithms are applied. These algorithms are divided into translational and rotational motion compensation algorithms. Translational motion compensation has two steps: range bin alignment and phase compensation. Rotational motion compensation also has two steps: removing MTRC (Migration Through Resolution Cells) and phase compensation. Compensation algorithms have high computational load. So sometimes other algorithms named image formation algorithms are used for imaging. In these algorithms, fourier transform is replaced by some other transforms like FRFT (FRactional Fourier Transform) or time-frequency transforms. These algorithms have less computational load but images which are produced in this way have less quality than that for the compensated images. One of the most useful criteria in both translational and rotational motion compensation is entropy. Entropy determines the degree of image quality. So it is proper in ISAR image compensation. In this thesis we introduce a new algorithm for rotational motion compensation by using entropy criterion and compare the results of this algorithm with some of image formation and compensation algorithms. We compare the results using normalized correlation parameter. This parameter computes the correlation between the result and the original image. The results show the effectiveness and robustness of the proposed algorithm. Keywords Imaging radars, ISAR, Focusing, Compensation, Fourier transform, Entropy.