<p> Non-coding RNAs (ncRNAs)—including microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and related species—have emerged as promising molecular biomarkers in epilepsy, but their diagnostic, prognostic, and therapeutic roles remain incompletely defined. We conducted a PRISMA 2020–compliant systematic review and meta-analysis (PROSPERO-registered). PubMed, Scopus, and Web of Science (2000–2025) were searched for studies evaluating ncRNA expression or clinical relevance in human epilepsy or validated animal models that reported diagnostic, prognostic, or treatment-response outcomes. Study quality was assessed with the Newcastle–Ottawa Scale, and pooled diagnostic estimates were generated when data allowed. Ninety-four studies met inclusion criteria: 79 investigated miRNAs, 10 lncRNAs, and several assessed piRNAs or circRNAs. The pooled area under the curve (AUC) for miRNA-based diagnostic models was 0.85, with mean sensitivity 76.5% and specificity 78.1%. lncRNAs showed a comparable mean AUC of 0.87, although based on fewer datasets. Recurrent biomarkers—particularly miR-134, miR-146a, and the lncRNA MALAT1—were consistently linked to drug-resistant epilepsy, higher seizure frequency, and poorer clinical control. Mechanistic analyses indicated that these ncRNAs regulate neuroinflammatory signaling, synaptic remodeling, and ion-channel–related pathways that are central to epileptogenesis. Overall, most studies were of moderate-to-high methodological quality. ncRNAs, especially circulating miRNAs, show strong potential as minimally invasive diagnostic and prognostic biomarkers in epilepsy and could support molecularly guided patient stratification and treatment monitoring. Future work should prioritize large prospective cohorts, assay standardization, and cross-platform validation to enable clinical translation.&#xa0;</p>