拓展常规方法到三参数和Ｌａｎｇｅｒ模型，提出了估计三参数、Ｌａｎｇｅｒ和 Ｂａｓｑｕｉｎ三种常用应力一寿命模型 Ｐ─Ｓ─Ｎ曲线及置信限的一种统一常规方法。方法将曲线表示为对数疲劳寿命均值和均方差线的广义形式，至多６个材料常数，并方便地采用线性回归技术求解。４５碳钢缺口试样（ｋｔ＝２）对称循环加载模式成组法疲劳试验结果分析说明了方法的有效性。三参数模型的拟合效果最好，Ｌａｎｇｅｒ模型稍差，Ｂａｓｑｕｉｎ模型较差。从拟合误差和安全性角度，Ｂａｓｑｕｉｎ模型不宜描述该套数据。此外，常规方法的估计结果可能受到试验数据局部统计特征参量的影响，要避免有必要遵循各加载水平试样数相同的原则，或发展改进的可以最大限度减小这种影响的方法。 Extending the conventional methods to the three-parameter and Langer models, a uniform conventional method for estimating P ─ S ─N curves and confidence limits of three commonly used stress-lifetime models of three parameters, Langer and Basquin was proposed. The method expresses the curve as a generalized form of logarithmic fatigue life means and mean square deviation, up to 6 material constants, and is conveniently solved using linear regression techniques. 45 Carbon Steel Notched Specimen (kt = 2) Symmetric Cyclic Loading Mode The fatigue test results of the group method demonstrate the effectiveness of the method. The fitting effect of the three-parameter model is the best, Langer model is slightly worse, Basquin model is poor. The Basquin model should not describe the set of data from the perspective of fitting error and safety. In addition, estimates of conventional methods may be affected by local statistical parameters of the test data to avoid the need to follow the same principle for the number of loading levels or to develop improved methods to minimize this effect.