This paper introduces an MSK non-coherent demodulation algorithm that integrates the soft spreading spectrum of CCSK, addressing the issue of signal demodulation in noisy environments. MSK modulation is widely used due to its efficient spectral utilization and low bit error rate, but traditional coherent demodulation relies on accurate carrier phase information, which is not suitable for short-burst communication and high-dynamic environments. The algorithm combines CCSK spreading technology, using base functions with strong periodic auto-correlation characteristics and their cyclic shift sequences, and proposes a waveform-matching-based MSK non-coherent demodulation algorithm, which significantly enhances the anti-interference capability, simplifies the receiver design, and maintains good demodulation performance under low signal-to-noise ratio conditions, with a high degree of tolerance to the sampling point drift. The simulation results show that under a signal-to-noise ratio of -5.5 dB, the system's bit error rate (BER) can be kept below 10^-6, demonstrating excellent noise resistance. Additionally, simulation experiments analyzed the impact of sampling point drift, revealing that a 1-sample point drift can cause 2-3 dB decrease in demodulation performance, while a 2-sample point drift can contribute to a 5-6 dB decrease, yet the demodulation performance remains within an operational range. The algorithm demonstrates outstanding demodulation performance under low signal-to-noise ratios and in high-dynamic scenarios. In conclusion, the waveform-matching-based MSK non-coherent demodulation algorithm proposed in this paper offers an efficient and reliable solution for wireless communication in complex environments.