In plant species, typical gene mapping strategies use populations initiated from crosses between two inbred lines. However, schemes including more than two parents could be used. In this paper, a new approach is introduced which uses a four-way cross population derived from four inbred lines. The four-way cross design for mapping quantitative trait loci (QTLs) provides tests for QTL segregation in four lines simultaneously in one experiment. Therefore, it is a more economical strategy than oneusing line crosses between only two lines. The new strategy also increases the probability of detecting QTLs if they segregate in one line cross but not in the other. A multiple linear regression analysis is used for QTL detection. It is proven that the expected residual variance from the regression analysis differs from the pure environmental variance. Correction for the bias is proposed and verified by computer simulations.