The potential of association mapping (AM) and genomic selection (GS) has not yet been explored for investigating resistance to soybean cyst nematode (SCN), the most destructive pest affecting soybean. We genotyped 282 representative accessions from the University of Minnesota soybean breeding program using a genome-wide panel of 1536 single nucleotide polymorphism (SNP) markers and evaluated plant responses to SCN HG type 0. After adjusting for population structure, AM detected significant signals at two loci corresponding to rhg1 and FGAM1 plus a third locus located at the opposite end of chromosome 18. Our analysis also identified a discontinuous long-range haplotype of over 600 kb around rhg1 locus associated with resistance to SCN HG type 0. The same phenotypic and genotypic datasets were then used to access GS accuracy for prediction of SCN resistance in the presence of major genes through a sixfold cross-validation study. Genomic selection using the full marker set produced average prediction accuracy ranging from 0.59 to 0.67 for SCN resistance, significantly more accurate than marker-assisted selection (MAS) strategies using two rhg1- associated DNA makers. Reducing the number of markers to 288 SNPs in the GS training population had little effect on genomic prediction accuracy. This study demonstrates that AM can be an effective genomic tool for identifying genes of interest in diverse germplasm. The results also indicate that improved MAS and GS can enhance breeding efficiency for SCN resistance in existing soybean improvement programs.