With biological and economic importance, threshold traits are discrete in phenotype but have the same polygenic genetic basis as quantitative traits. The traditional linkage analysis for quantitative traits is invalid for threshold traits due to their special characters. The transmission disequilibrium test (TDT) has received great attention recently in localizing human disease genes due to its simplicity and powerfulness. But TDT only deals with data from independent nuclear families and it will lose information about extended pedigree that incorporates information not only from parents and siblings but also from other relatives. The pedigree disequilibrium test (PDT) proposed by Martin in 2001 can be used to analyze the extended pedigree in human. In this study, PDT was introduced into the QTL mapping of threshold traits for farm animals, and was modified in order to accommodate the pedigree structures of farm animals. The modified PDT was renamed pedigree transmission disequilibrium test (PTDT) and its power and type I error were investigated and compared with that of PDT by Monte Carlo simulation. It was shown that PTDT is a robust and valid approach to mapping QTL of threshold trait. When the parental information is complete, PTDT and PDT are almost the same in terms of power and type I error. However, if the parental information is missing to a certain extent, PTDT is higher in power and lower in type I error than PDT. These results imply that PTDT can be a novel approach to QTL fine mapping of threshold traits based on the existing coarse mapping information.