This paper addresses the issue of finding an efficient estimator of the normal populations variance using the sample estimate of the population coefficient of variation. The proposed estimator is arrived at by using the well-known Searls (1964)’ perspective for the MMSE estimation of the normal population mean. Recently, Searls & Intarapanich (1990) found the MMSE of the normal population variance. We have presented the relative efficiency of our estimator with respect to usual unbiased estimator S2 vis-à-vis that of the Searls & Intarapanich (1990)’s estimator by means of an empirical simulation study.