This paper conducts an in-depth analysis of the noise generation mechanism in MEMS gyroscopes, introduces the Allan variance analysis method and its application in MEMS gyroscope error analysis. Addressing the limitations of the classical Allan variance analysis method in MEMS gyroscope data analysis, an optimized noise analysis method based on Allan variance for MEMS gyroscopes is proposed, and its effectiveness is verified through simulation experiments. This provides a theoretical foundation for better understanding and analyzing the random error characteristics of MEMS gyroscopes, thereby enhancing their performance in various applications.