提出了一种改进的添加自适应高频谐波进行经验模态分解(EMD)的算法，减少了EMD原始算法中频谱混叠现象。通过对原始信号的预处理，自动提取出原始信号中包含的最高频率分量，并根据提取出的频率分量进行高频谐波添加。仿真验证了添加自适应高频谐波的EMD 算法，可有效减少EMD 算法中频谱混叠现象，同时解决了高频谐波添加中频率难以确定的问题。

An improved Empirical Mode Decomposition(EMD) method adding high frequency harmonic wave with self-adaption is put forward, aiming to address the problem of the mode aliasing in the original empirical mode decomposition algorithm. The method tries to extract the highest frequency component of the original signal and adds it to the original signal. The simulation shows that the improved EMD algorithm could solve both the mode aliasing problem and the problem that the frequency of adding wave is difficult to determine.