Abstract: Under the given confidence level, confidence interval estimate of the fatigue life is an important part of the stress-life relational curve design. Aiming at the problem of confidence interval of the fatigue life, this paper introduces three kinds of commonly used solving methods which include the method of mathematical statistics, the MATLAB software solving method and bilateral confidence interval. At the same time, the confidence interval of a group of fatigue test data was solved by using the linear model and three-parameter model of the stress life curve. By analyzing and comparing their calculation processes and results, it's concluded that when solving the logarithmic life confidence interval by the regression equation, the estimation precision of the interval increases with the increase of sample size and the logarithmic fatigue life.For the confidence interval estimate of a middle or long life area, three parameter model is better than the linear model, it's feasible to use MATLAB software solving method instead of the mathematical statistics method to solve the confidence interval.