Illinois long‐term selection strains of maize (Zea mays L.) have been useful for identifying genomic regions controlling kernel oil, protein, and starch concentrations. To identify kernel trait quantitative trait loci (QTL) in a genetic background more relevant to practical breeding, 150 BC1‐derived S1 lines (BC1S1s) were produced from Illinois High Oil and recurrent parent B73. Oil, protein, and starch were measured in BC1S1s and in Mo17‐topcross hybrids (TCs). Kernel mass of BC1S1s and grain yield of TCs were also determined. Starch was positively correlated with mass in BC1S1s (rp = 0.67**, α ≤ 0.01) and with yield in TCs (rp = 0.59**). Oil was negatively correlated with mass in BC1S1s (rp = −0.29**) and with yield in TCs (rp = −0.30**). Oil was negatively correlated with starch in BC1S1s (rp = −0.75**) and TCs (rp = −0.66**). A genetic map with length = 1486 cM was created with 110 markers. Multiple regression models with QTL detected by composite interval mapping (CIM) explained 46.9, 45.2, 44.3, and 17.7% of phenotypic variance for oil, protein, starch, and mass, respectively, in BC1S1s and 17.5, 22.9, 40.1, and 28.7% for oil, protein, starch, and yield, respectively, in TCs. A 22 cM‐interval on chromosome 6 in BC1S1s included oil, protein, and starch QTL, including a QTL explaining 36.7% of the BC1S1 phenotypic variation for oil. No yield QTL were detected in this region. Introgression of this QTL into breeding lines might increase oil while maintaining yield.