This article addresses the problem of estimating a family of general population parameter ${ heta}_{({alpha},{eta})}$ using auxiliary information in the presence of measurement errors. The general results are then applied to estimate the coefficient of variation $C_Y$ of the study variable Y using the knowledge of the error variance ${sigma}^2{_U}$ associated with the study variable Y, Based on large sample approximation, the optimal conditions are obtained and the situations are identified under which the proposed class of estimators would be better than conventional estimator. Application of the main result to bivariate normal population is illustrated.