Recently, using human behavior to hide the existence of information has been at the center of steganography research. In this study, a behavioral steganography algorithm using CMI (Coded Signal Inversion) coding is proposed to minimize the high bit error rate that occurs when transmitting a large number of continuous and identical confidential information in the knapsack algorithm, which is used to improve information transmission efficiency and flexibility of transmission mode in social networks. In the proposed algorithm; Data redundancy is reduced by reducing the number of mutual friends of the sender and each receiver. Then, the proposed algorithm was applied and the results were analyzed. Experimental analysis shows that this scheme improves the practical value of behavioral steganography in social networks and has high security.