As a prerequisite for extracellular vesicle (EV) -based studies and diagnosis, effective isolation, enrichment and retrieval of EV biomarkers are crucial to subsequent analyses, such as miRNA-based liquid biopsy for non-small-cell lung cancer (NSCLC). However, most conventional approaches for EV isolation suffer from lengthy procedure, high cost, and intense labor. Herein, we introduce the digital microfluidic (DMF) technology to EV pretreatment protocols and demonstrate a rapid and fully automated sample preparation platform for clinical tumor liquid biopsy. Combining a reusable DMF chip technique with a low-cost EV isolation and miRNA preparation protocol, the platform completes automated sample processing in 20-30 min, supporting immediate RT-qPCR analyses on EV-derived miRNAs (EV-miRNAs). The utility and reliability of the platform was validated via clinical sample processing for EV-miRNA detection. With 23 tumor and 20 non-tumor clinical plasma samples, we concluded that EV-miR-486-5p and miR-21-5p are effective biomarkers for NSCLC with a small sample volumn (20-40 μL). The result was consistent to that of a commercial exosome miRNA extraction kit. These results demonstrate the effectiveness of DMF in EV pretreatment for miRNA detection, providing a facile solution to EV isolation for liquid biopsy.