On the basis of a high-density genetic map with a large number of molecular markers quantitative trait loci (QTLs) can be finely mapped. However, the information of a dense map cannot be fully utilized by early segregating generations such as an F₂ and a backcross due to insufficient number of recombination events among closely linked markers. One needs to advance generations to increase the number of recombination events in order to take advantage of a dense map. In this paper the utility of populations in advanced generations such as F₅, F₆ or recombinant inbred lines derived from F₂ by repeated cycles of self-fertilization in mapping of QTLs is discussed. The effectiveness of mapping a QTL based on advanced generations is evaluated in a simple regression model. The power of detection of a QTL is enhanced and the lengths of support intervals of map location of a QTL are reduced as generations advances, especially the efficiency of detecting and locating a QTL is significantly improved in F₃-F₅ in comparison with F₂. Moreover a multiple regression model based on a large number of markers is devised to separate two linked QTLs. The ability of dissolving multiple linked QTLs is increased as generations advance.