To address the limited transmission capability of power grid monitoring information in remote and disaster?affected areas, this paper proposes a Low Earth Orbit(LEO) satellite?based Internet of Things(IoT) transmission scheme tailored for emergency power grid monitoring applications. Firstly, a multi?level priority scheduling mechanism is developed by incorporating key transmission parameters such as the number of retransmissions and backoff time, thereby ensuring the accuracy and timeliness of information delivery according to different emergency levels. Secondly, in consideration of the highly dynamic characteristics of satellite/Unmanned Aerial Vehicle(UAV) communication channels, a closed?loop demodulation method is proposed to mitigate the impact of time?varying Doppler shifts. This method extracts dynamic Doppler frequency offset information by analyzing the relationship between the main and side peaks in Fast Fourier Transform(FFT)?based demodulation decisions, and introduces a feedback loop to dynamically compensate for the frequency offset in subsequent symbol demodulation. This significantly enhances the Bit Error Rate(BER) performance of IoT transmissions under high?mobility channel conditions. Finally, to validate the effectiveness of the proposed method, a comprehensive performance evaluation is conducted through Matlab simulations and Field?Programmable Gate Array(FPGA)?based hardware testing. Both simulation and FPGA results demonstrate that the proposed closed?loop demodulation scheme achieves a markedly better BER performance than the traditional method. At a BER of 10^?3, it provides an Signal-to-Noise Ratio(SNR) gain of about 2 dB at Spreading Factor(SF)=9, and the BER curve no longer shows a flooring effect as the SNR further increases.