On the Internet, stego images are often transmitted under blind compression, which significantly limits the robustness of secret information extraction. This paper proposes a novel image hiding method, called IHCP, which proactively predicts the quality factor to enhance robustness against blind JPEG compression and lossy transmission. In IHCP, the Decoder Guidance Module (DGM) predicts the quality factor of the stego image, providing guidance to the decoder for more accurate extraction of secret information from stego images subjected to blind compression. The Feature Fusion Decoder (FFD) combines the stego image with the extraction results from the invertible decoder, guided by the DGM, to extract finer details of the secret information. Finally, we employ a Siamese network to align the features of the compressed stego image with those of the corresponding uncompressed stego image, further improving robustness. Comprehensive experiments on real datasets demonstrate that the proposed method significantly outperforms baseline models in resisting JPEG compression and is robust to lossy transmission in social networks.