Genome‐wide association studies (GWASs) have identified single nucleotide polymorphisms (SNPs) associated with susceptibility and severity of coronavirus disease 2019 (COVID‐19). However, identified SNPs are inconsistent across studies, and there is no compelling consensus that COVID‐19 status is determined by genetic factors. Here, we conducted a systematic review and meta‐analysis to determine the effect of genetic factors on COVID‐19. A random‐effect meta‐analysis was performed to estimate pooled odds ratios (ORs) of SNP effects, and SNP‐based heritability (SNP‐h2) of COVID‐19. The analyses were performed using meta‐R package, and Stata version 17. The meta‐analysis included a total of 96,817 COVID‐19 cases and 6,414,916 negative controls. The meta‐analysis showed that a cluster of highly correlated 9 SNPs (R2 > 0.9) at 3p21.31 gene locus covering LZTFL1 and SLC6A20 genes was significantly associated with COVID‐19 severity, with a pooled OR of 1.8 [1.5–2.0]. Meanwhile, another 3 SNPs (rs2531743‐G, rs2271616‐T, and rs73062389‐A) within the locus was associated with COVID‐19 susceptibility, with pooled estimates of 0.95 [0.93–0.96], 1.23 [1.19–1.27] and 1.15 [1.13–1.17], respectively. Interestingly, SNPs associated with susceptibility and SNPs associated with severity in this locus are in linkage equilibrium (R2 < 0.026). The SNP‐h2 on the liability scale for severity and susceptibility was estimated at 7.6% (Se = 3.2%) and 4.6% (Se = 1.5%), respectively. Genetic factors contribute to COVID‐19 susceptibility and severity. In the 3p21.31 locus, SNPs that are associated with susceptibility are not in linkage disequilibrium (LD) with SNPs that are associated with severity, indicating within‐locus heterogeneity.