Three-dimensional (3D) target reconstruction provides a more specific description of target features, which would be beneficial for the classification process. In this paper a 3D imaging technique of non-cooperative target acquired by airborne radar system is presented. Traditional two-dimensional (2D) single channel inverse SAR (ISAR) imaging locates each target scattering center in range and Doppler. The third point scatterer coordinate that allows to reconstruct the target in a 3D space is estimated using interferometric phases of two non-parallel baselines and Doppler point scatterer location. The accuracy of the 3D reconstruction depends on the length of the baselines and on the focusing achieved by the 2D imaging process for each channel. The relative movement between the target and the airborne system increases the difficulty to form a well-focused 2D image. The radial component of both radar and target motion affects the image focusing. Moreover traditional image formation method as range/Doppler algorithm introduces further blurring due to the wideband situation. To improve the image quality resulting from standard radial motion compensation techniques and range/Doppler image formation, a method for 2D image processing based on range/velocity data formatting is additionally applied.