In order to achieve a high range resolution and low instantaneous bandwidth, stepped frequency synthetic aperture radar (SAR) is widely used with stepped-frequency chirps and frequency-domain spectrum reconstruction method. To combine the spectrum reconstruction with SAR image algorithms, “Image for each sub-band and then reconstruct the wideband spectrum” is a more adaptable flow since its validity for almost all imaging algorithms. However, due to the two-dimensional (2-D) spectrum twisting, a special problem called range spectrum aliasing may occur and lead the flow to invalidity, especially in high-resolution SAR systems. In this paper, we discuss the extended range spectrum by the analysis of the 2-D twisting spectrum, judge its influence through an azimuth threshold resolution, and propose an improved stepped-frequency SAR imaging algorithm with the range spectral-length extension (RSLE) strategy, which can eliminate the range spectrum aliasing. After theoretic analyses, simulations and experimental results are provided to demonstrate the proposed algorithm.