设备诊断技术是先进生产力的重要组成部分的理念正在被越来越多的人所接受。为了深入开展简易诊断工作,本文结合工作实际介绍了应用数理统计分析的重要方法—最小二乘估计对振动测试数据进行线性回归,实施趋势分析和劣化倾向管理的方法,指出了在具体应用中应注意的问题,绐出了常见的非线性回归化为线性回归的方法。实践证明,该方法对于预知维修工作有着十分重要的指导意义。随着计算机技术应用的普及,数理统计分析方法必将在简易诊断工作中显示出强大的威力。 The idea that device diagnostics is an important part of advanced productivity is being accepted by more and more people. In order to deeply carry out the simple diagnosis work, this paper introduces the important method of applying mathematical statistics and analysis - Least Squares Estimation to linear regression of vibration test data, the trend analysis and the method of deterioration propensity management, and points out that in practical application Attention problems, out of the common method of nonlinear regression to linear regression. Practice has proved that this method has very important guiding significance for predicting maintenance work. With the popularization of the application of computer technology, mathematical statistics and analysis methods will certainly show great power in the simple diagnosis.