In this article, a new estimator for the population mean at the current occasion in successive sampling having two occasions is proposed under the missing data case. The minimum mean square error and optimum fraction of a fresh sample of the proposed estimator are obtained. This estimator is compared with the estimator suggested by Singh and Priyanka (Singh, G. N., Priyanka, K. Effect of non-response on current occasion in search of good rotation patterns on successive occasions, Statistics in Transition 8 (2), 273-292, 2007), and the efficient condition for the proposed estimator is found. In a numerical example, the mean square errors of these estimators are also computed according to various values for the fraction of missing data, coefficient of correlation, fraction of sub sampling and fraction of fresh sample. The results of the numerical example show that the proposed estimator is more efficient than Singh-Priyanka estimator for all values of the parameters.