Abstract: Highly squinted airborne synthetic aperture radar (SAR) with a considerable flexibility has capability of target imaging in advance and repeatedly. Nevertheless, high-precision imaging algorithms for high squint SAR are generally complicated and have a great computational load that cannot meet the requirement on real time. In this paper, a real-time imaging processing architecture based on multi-core digital signal processor (DSP) is proposed, which is built on the polar format algorithm. First, two-cascade smooth filtering and azimuth down-sample operations are employed for processing the SAR echo data. Then, the principle of chirp scaling is exploited to the range resampling, into which coarse correction method of range cell migration is incorporated. As a consequence, the maximum number of time-domain intercept points and the reduction of computational load in subsequent processing are accomplished, which will be cascaded with high-precision sinc-based interpolation to realize the azimuth resampling. Finally, autofocus technique is employed for further compensating the azimuth phase error so as to obtain the highly focused SAR imagery. Experimental results using the measured high squint airborne SAR data indicate that the proposed DSP processing architecture is capable of processing the 10 K×4 K pixel points in azimuth and range within 0. 873 s to generate 2 K×2 K imagery, and hence it can satisfy the requirement of real-time imaging design.