AbstractThis paper discusses the asymptotic representations of a class of L2-distance estimators based on weighted empirical processes in a multiple linear regression model when the errors are a function of stationary Gaussian random variables that are long-range dependent.