A method for high-throughput isolating quantitative trait controlling genes, comprising following steps : 1) constructing target trait isolated population : the population is one of F2, F3, and double haploid (DH) system and recombinant inbred line (RIL), which are offspring of two parents hybrid; 2) mixing extreme samples in the population and isolating total RNA : dividing the isolated offspring population into three types according to phenotype of the trait, i.e. extremely strong phenotype of the trait, middle phenotype of the trait and extremely weak phenotype of the trait, and then mixing same amount of samples with extremely strong phenotype of the trait or samples with extremely weak phenotype of the trait into two extreme population mixing samples; 3) analyzing gene expression : comparing the difference of gene expressions between two extreme mixing samples by gene expression analysis method such as chip, expressed sequence tag (EST) sequencing, suppression subtractive hybridization(SSH), cDNA-amplified fragment length polymorphism (AFLP) and so on; and 4) obtaining and verifying candidate gene : finding out the genes which show different expression between the two extreme mixing samples to be the candidate controlling genes related to the target trait, verifying the functions of these genes by expressing the genes with transgenic method, molecule labeling relativity analysis, and contribution rate analysis, and obtaining the target gene which can control phenotype of the trait. The method is useful for isolating genes for controlling quantitative trait which is controlled by several genes in all species, which is simple, rapid, high-throughput and economic. A gene for controlling oil content of Brassica napus, which is regarded as quantitative trait, is isolated. The nucleotide sequence of the candidate gene HD1 is provided and the function thereof in improving oil content of Brassica napus seed is also verified.