In models related to economics and business sciences, more often there are some limitations on the variables due to certain conditions. These limitations cause fitting mathematical models to real situations and developing economical methods. In these models, the ranges of the dependent variables are restricted to some subsets of the real line. In this thesis, we investigate censored regression model with specified censored with bound zero. At first this model was proposed by Tobin (1958) and called Tobit model. Powell in 1958 introduced estimators based on regression quantiles for censored regression model. This model called censored quintile regression. In this thesis we investigate censored LAD estimators that are certain form of Powell's estimator. Powell showed that censored LAD estimator had asymptotically normal distribution and for calculating variance-covariance matrices of this estimator we need to know error distribution. The asymptotic variance-covariance matrices are difficult to estimate since they involve conditional densities of error term. One of the possible methods for estimating distribution of Powell’s regression quantile estimators for censored model and doing statistical inference is Bootstrap method. In this thesis we introduce efficient computational methods for approximating distribution of Powell’s estimator between certain resampling methods. Then we propose empirical application of quantile regression and at the end we propose the simulation study for investigating achievement of modified estimators.