This paper considers the problem of determining the optimal process mean and screening specification limits of a surrogate variable associated with product quality under two-stage screening procedure. In two-stage screening, the surrogate variable is inspected first to decide whether an item should be accepted, rejected or additional observations should be taken. If additional observations are required, the performance variable of interest is then observed to classify the undecided items. Assuming that the performance variable and the surrogate variable are jointly normally distributed, the optimal process mean and the screening limits are obtained by maximizing the expected profit which includes selling price, production, reprocessing, inspection and penalty costs. A numerical example is presented and numerical studies are performed to compare the proposed two-stage screening procedure with single-stage screening procedures.