Soybean, a source of plant-derived lipids, contains an array of fatty acids essential for health. A comprehensive understanding of the fatty acid profiles in soybean is crucial for enhancing cultivars and augmenting their qualitative attributes. Here, 180 F10 RILs, derived from the cross-breeding of ‘Jidou 12’ and ‘Y9’, were used as primary experimental subjects. Concurrently, a GWAS was conducted on 290 samples from a natural population to scrutinize the five fatty acid components for three consecutive years. The QTL analysis elucidated 23 stable QTLs intricately associated with the five fatty acid components, exhibiting phenotypic contribution rates ranging from 2.78% to 25.37%. In addition, GWAS of the natural population unveiled 102 significant loci associated with these fatty acid components. The haplotype analysis of the colocalized loci revealed that Glyma.06G221400 on chromosome 6 exhibited a significant correlation with stearic acid content, with Hap1 showing a markedly elevated stearic acid level compared with Hap2 and Hap3. Similarly, Glyma.12G075100 on chromosome 12 was significantly associated with the contents of oleic, linoleic, and linolenic acids, suggesting its involvement in fatty acid biosynthesis. Furthermore, genomic selection (GS) for fatty acid components was conducted using ridge regression best linear unbiased prediction based on both random SNPs and SNPs significantly associated with fatty acid components identified by GWAS. GS accuracy was contingent upon the SNP set used. Notably, GS efficiency was enhanced when using SNPs derived from QTL mapping analysis and GWAS compared with random SNPs, and reached a plateau when the number of SNP markers exceeded 3,000.