Positive-strand RNA viruses remodel host endomembranes to form replication organelles (ROs). However, the membrane sources for coronavirus double-membrane vesicle (DMV) biogenesis remain unclear. Here, we used proteomic, genetic, pharmacological, and imaging approaches to identify the central membrane supply route for DMV formation. APEX2-based proximity proteomics first identified candidate pathways, followed by functional validation using siRNA knockdown, CRISPR/Cas9 editing, and Golgicide A (GCA) inhibition. DMV formation and viral RNA synthesis were assessed by electron microscopy, immunofluorescence, and qRT-PCR. Golgi–DMV association was analyzed by sucrose gradient fractionation and confocal imaging. In vivo relevance was evaluated in C57BL/6J mice infected with mouse hepatitis virus (MHV) and treated with GCA (intraperitoneal, 50 mg/kg; n = 4–5 per group). We showed that coronaviruses hijack the host GBF1-ARF1-COPI machinery to redirect Golgi-derived membranes to viral replication sites for DMV formation and expansion. Disruption of this pathway markedly impairs viral RNA synthesis, DMV biogenesis, and replication of multiple coronaviruses, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Pharmacological inhibition of Golgi-specific Brefeldin A-resistance factor 1 (GBF1) suppressed viral replication and pathology in vivo. Together, our findings reveal a previously unrecognized Golgi-derived membrane supply route for coronavirus ROs and highlight the GBF1-ARF1-COPI pathway as a potential host-directed target for broad-spectrum antiviral therapies.