This article proposes a novel behavior steganography model based on secret sharing, the main idea of which is to use secret messages as random elements in the secret sharing process to generate shadow images. Based on the introduced model and analyzing two secret image sharing algorithms — threshold secret image sharing (SIS) and threshold visual secret sharing (VSS), two specific behavior steganography schemes are presented, which are implemented by utilizing secret sharing behavior. In the embedding phase, the random selection behavior is employed to hide secret messages. In the extraction phase, when the secret image is recovered from shadow images, secret messages can also be extracted successfully. The contribution of the authors model is that two secret information transmission channels are opened, which provides a large amount of hidden capacity and has loss tolerance and so on. Experimental results and analyses demonstrate the effectiveness of the proposed scheme. It has both good imperceptibility and large capacity, but the robustness of their scheme is poor.