为实现精神疲劳状态的客观量化连续监测，本研究基于心率变异性的特征信息，研究了不同疲劳状态的量化识别方法。设计实验采集了不同疲劳状态下的心电信号，采用小波变换去除噪声干扰，提取了心率变异性的时域、频域等特征参数，并对所提取的特征参数对比分析以构建支持向量机等多种分类模型，最终利用支持向量机算法实现了正常、疲劳、嗜睡三种不同程度的精神状态分类。结果表明，本方法对于疲劳状态识别具有较高准确性，通过优化支持向量机的参数，对于不同疲劳状态的识别精度可超过80%。将此技术运用到日常生活中，可对人的主观疲劳状态进行客观化评价，在交通、教育、医学监护等方面具有重要的研究意义和应用前景。 In order to achieve the objective quantitative continuous monitoring of mental fatigue state, the quantitative identification method of different fatigue states is studied based on the characteristic information of heart rate variability. The ECG signals under different learning duration were col-lected in the design experiment. Wavelet transform was used to reduce the noises or interference, and characteristic parameters of heart rate variability (HRV) were extracted in the time domain and frequency domain. The extracted characteristic parameters were compared and analyzed to construct various classification models such as support vector machine. Finally, three mental states of normal, fatigue and drowsiness were classified by support vector machine algorithm. The results show that this method has a high accuracy for fatigue state identification, and the identification ac-curacy of different fatigue state can be more than 80% by optimizing the parameters of support vector machine. The application of this technology in daily life can objectively evaluate people’s subjective fatigue state, which has important research significance and application prospect in transportation, education, medical monitoring and so on.