The complex genetic architecture of heritability in multiple sclerosis (MS) remains undisclosed mainly. Epistasis (gene-gene interaction) substantially impacts MS; however, it is largely unexplored, especially among the non-coding RNA genes and their targets. The long non-coding RNA GAS5 exacerbates demyelination and sponges miR-146a and miR-155, impeccable contributors to MS pathogenesis. miR-146a negatively regulates the immune responses by targeting IRAK-1. We investigated the association of epistatic effects and haplotypes of five single nucleotide polymorphisms (SNPs), GAS5 rs2067079, miR-146a rs2910164 and rs57095329, IRAK-1 rs3027898, and miR-155 rs767649, with the risk of MS and its phenotypes. The expression quantitative trait locus (eQTL) associated with these variants was explored through bioinformatics analysis. The study enrolled 116 MS patients and 120 healthy controls. No strong linkage disequilibrium (D' ≥ 0.8) was detected among the studied SNPs. SNP-SNP interactions overlaid an overall magnified risk of MS and its phenotypes compared to the single-locus effects. After adjustment for multiple comparisons, the most significant interactions associated with the risk of overall MS and secondary-progressive MS were rs2067079-rs2910164, rs2910164-rs57095329, and rs3027898-rs767649. The last two former SNP-SNP interactions were highly associated with relapsing-remitting MS risk. The same pattern of interactions, as observed in association with MS risk, was female-specific. The CCAAA haplotype (alleles in the order of rs2067079, rs2910164, rs57095329, rs3027898, and rs767649) was protective against MS risk (CCAAA vs. CGAAT, adjusted OR = 0.14, 95% CI = 0.03-0.69, P = 0.009). Among MS patients, harboring the CGACT and CGAAT haplotypes was more prevalent in females and males, respectively. MS patients having EDSS ≥ 6 had a significantly higher frequency of the CCGCA haplotype than those with EDSS < 6. Functional analysis revealed rs2067079, rs57095329, and rs767649 as strong cis-eQTL regulating multiple genes, particularly in the brain and immune system. We propose that a magnified combined effect of GAS5, miR-146a, IRAK-1, and miR-155 genetic variants via epistatic interactions might impact the risk of MS and its phenotypes and could help in the risk stratification of MS patients.