Analysis of the low-density lipoprotein receptor (LDLR) gene based on the rs688 and rs5925 genetic polymorphisms has provided evidence suggesting that haplotypes related to rs688 and rs5925 are associated with ischemic cerebrovascular diseases. Both rs688 and rs5925 have been empirically identified as exon-splicing enhancers in silico, and rs688 has been shown to be a functional polymorphism that modulates LDLR exon 12 splicing efficiency both in vitro and in vivo. The aim of this study was to evaluate whether rs688 and rs5925 and their haplotypes may alter the splicing efficiency of exons 12 and 13 both in vivo and in vitro. When the minigenes were evaluated for splicing efficiency, we found that converting rs688C to the T allele reduced exon 12 splicing efficiency. In parallel, converting rs688T to the C allele increased the efficiency of exon 12 inclusion. The apparent difference in splicing efficiency was 9.36%±2.58% between the C and T alleles. When rs688C existed in the minigene, the major and minor rs5925 alleles were also sufficient to account for the differences in splicing efficiency of LDLR involving exon 13. The apparent splicing efficiency difference was 5.43%±2.87%. Sequential mutations of rs688 and rs5925 were performed to generate four different haplotypes in the LDLR minigene system. The splicing efficiencies for the haplotypes CC, CT, TC, and TT were 79.60%±1.38%, 76.68%±0.85%, 69.02%±1.79%, and 68.54%±1.38%, respectively. The splicing efficiency of the four haplotype groups differed significantly. In vivo analysis of human leukocyte samples was also compatible with in vitro analysis, indicating a mutual effect between rs688 and rs5925 in regulating LDLR splicing efficiency.