Molecular markers such as single nucleotide polymorphisms (SNPs) and SilicoDArT are important in dissecting the genetic diversity of a population at DNA level. The two marker types were analyzed using the same genotyping platform. Although the two marker types were analyzed using the same genotyping platform, they were filtered using a different marker stringency. We determined the genetic diversity of 623 half-sibs of shea tree (Vitellaria paradoxa a C. F. Gaertn. nilotica) assembled from five geographical locations (Katakwi, Otuke, Amuru, Moyo, and Arua) in Uganda’s shea parklands. A total of 27,063 Diversity Arrays Technology (DArT) SNP and 9018 SilicoDArT markers were used to assess genetic diversity and population structure. The populations showed a low genetic diversity (Gst = 0.21), very low population differentiation (FST = 0.02), low-to-moderate linkage disequilibrium (r = 0.2), and Hardy–Weinberg Equilibrium (HWE = 0.1982). The study further revealed a higher genetic diversity within population (0.26) than among the population (0.21). A high level of heterozygosity was observed within individuals (0.26) and markers (0.32) revealing a high non-random association of alleles at different loci that offer opportunities to design association studies and allele transfer in marker-assisted selection in the population. The markers varied in their polymorphic information content values (SilicoDArT = 0.11) and (SNPs = 0.15) on genetic diversity. This study reveals the importance of genetic diversity and population structure analysis using high-density DArT-Seq SilicoDaRT and SNP makers in the conservation and breeding of shea tree in Uganda.