Synthetic aperture radar (SAR) systems produce large amounts of data which needs to be compressed and transferred. SPIHT compression algorithm performance superior, but the coded images have fragile visual quality when transmitted across the noisy channel. An errortolerance mechanism is necessary. An errorresilience algorithm called LL coefficients hiding (LLCH) for transmission of synthetic aperture radar (SAR) amplitude compression images in noisy channel is presented. The proposed method partitions data based on the coding process after the set partitioning in hierarchical tree (SPIHT’s) coding process, to ensure that the decoder can resynchronize after the error occurs. The most significant low frequency and sublow frequency coefficients are embedded into the end of high frequency data packets to achieve the purpose of protecting the important coefficients. For the high frequency coefficients of damage or loss, linear interpolation is used to recover the coefficients rely on the correlation between subbands. The experimental results show that the algorithm can effectively improve the quality of the restored image under the tiny redundancy cost.