The balance between trace minerals in rice, a staple food for more than half of the world’s population, is crucial for human health. However, the genetic basis underlying the correlation between trace minerals has not been fully elucidated. To address this issue, we first quantified the concentrations of 11 trace minerals in the grains of a diversity panel of 575 rice cultivars. We found that eight elements were accumulated at significantly different levels between the indica and japonica subspecies, and we also observed significant correlation patterns among a number of elements. Further, using a genome-wide association study, we identified a total of 96 significant association loci (SALs). The differentiation of the major-effect SALs along with the different number of high-concentration alleles present in the two subspecies shaped the different element performance in indica and japonica varieties. Only a few SALs located in clusters and the majority of SALs showed subspecies/subgroup differentiation, indicating that the correlations between elements in the diversity panel were mainly caused by genome differentiation instead of shared genetic basis. The genetic architecture unveiled in this study will facilitate improvement in breeding for trace mineral content.