Range-Doppler (RD) algorithms are widely used in inverse synthetic aperture radar (ISAR) imaging. In the standard RD algorithm, envelope alignment and autofocus are first applied to transform the original data into equivalent turntable target data, and then FFT is used for the image formation. Usually, the migration through resolution cells (MTRCs) due to target rotation is ignored. With the improvement of resolution or the increase of target size, MTRCs cannot be ignored and must be mitigated. For high speed moving targets, the most widely used 'stop-and-go' data model is violated. In this case, the mitigation of MTRCs becomes even more complicated. In the paper, technical issues associated with high resolution ISAR imaging of high speed moving targets are addressed. A preprocessing algorithm is first proposed to convert the non-coherent raw ISAR data into coherent data, then a keystone formatting algorithm developed for the imaging of slow ground moving targets in SAR is applied to mitigate the MTRCs owing to rotational motion, and, finally, a high resolution time-frequency analysis based range-instantaneous Doppler (RID) algorithm is used to produce the focused ISAR image. Numerical examples are provided to illustrate the performance of the proposed approach.