Cyber-energy dual modulation technology is emerging as the communication foundation for efficient distributed coordinated control of multi-source converters in DC microgrids. It offers advantages over traditional optical fiber and power line carrier communication by eliminating the need for dedicated lines and specific signal generation units. However, existing demodulation methods predominantly rely on Discrete Fourier Transform (DFT) for spectrum analysis and face challenges of spectral leakage and insufficient real-time performance under the "resource-limited, non-synchronous sampling" constraints of embedded systems. This paper proposes an efficient demodulation technology based on the Goertzel algorithm. Firstly, taking the typical power/data 2FSK single-carrier modulation as the research subject, it analyzes its demodulation challenges. Secondly, an efficient demodulation scheme based on the Goertzel algorithm is designed, which alleviates spectral leakage and real-time limitations with lower computational complexity. Finally, systematic simulation comparison experiments are conducted for validation. Results show that, under the same computational budget, the proposed method reduces the root-mean-square error of amplitude estimation by up to 98% compared to the conventional DFT-based scheme, effectively suppressing spectral leakage. Under equal analysis window length, its processing time is reduced by an average of 60.5% compared to the traditional DFT algorithm, achieving a speedup factor of 2.53.