This paper proposes a new estimator of the parameter sigma2 in simple
linear regression model when the observations are randomly censored on
the right. This estimator is explicitly defined and easily computable
. The strong consistency and the asymptotic normality are also establi
shed under certain conditions.