Hepatocellular carcinoma (HCC) is the second leading cause of cancer‐related death worldwide. China accounts for over half of the new cases and deaths. Diagnostic imprecision and a lack of complimentary molecular biomarkers are partially responsible for this lack of progress. Herein, serum‐derived exosomal microRNA (miRNA) profiling was performed on 80 patients which histologically confirmed HCC and 30 normal controls. A classification of 8 exosomal miRNAs had biologically and statistically significant differences between HCC and normal serum samples, including miR‐122, miR‐125b, miR‐145, miR‐192, miR‐194, miR‐29a, miR‐17‐5p, and miR‐106a. Online algorithm showed strong independent classification accuracy (area under the curve) reached 0.535 to 0.850, separately. The significant correlation between serum exosomal miRNAs and tumor size was observed. In addition, the survival difference of HCC patients with high or low exosomal miR‐106a was statistically significant using Kaplan‐Meier analysis. Besides, we also measured the proliferation and invasion ability of HCC cells following exosomal miR‐106a mimics or inhibitor treatment. After prediction with algorithms, mitogen‐activated protein kinase and c‐Jun N‐terminal kinase pathways were identified associated with miR‐106a’s function. In summary, differentially expressed serum exosomal miRNAs can be helpful for diagnostic and prognostic of HCC.