Weibull distribution is one of the most commonly used lifetime distribution. Its hazard rate is a flexible one in modeling failure time data, in the sense that it can be increasing, constant or decreasing. Many authors have studied its characteristics and have proposed powerful and widely used estimator of its parameters. In general, theses estimators assume that the sample size is large and / or the observations include no outlier. However, real reliability – data often contradict these assumptions. In this thesis, we use the bootstrap method for estimation of parameters of weibull distribution with two parameters in the presence of outlier.