A transmission method based on trusted cryptographic modules is proposed for secure information transmission on the edge side of the power IoT. The SG-Edge trusted edge computing framework is constructed to optimize the execution efficiency of information transmission under the constraint of ensuring the security of transmission tasks. By combining USBKey dual-factor and TCM authentication mechanisms, the security of the information transmission channel is measured through quantitative processing. Through multiple rounds of composite encryption process inverse operations, a trusted key is generated to ensure the reciprocity of wireless channels. The initial coherence time is calculated so that the adjusted training sequence length can complete secure data transmission within the coherence time period. The experimental results show that the method has errors of up to 10, 4, and 5 A between the three-phase current and actual data during the time period from 12:00 to 21:00. Compared with other methods, the transmission error is minimal, indicating that the use of this method can protect the security of information transmission on the edge side.