The profiling of cancer cell secretomes is considered to be a good strategy for identifying cancer-related biomarkers, but few studies have focused on identifying low-molecular-mass (LMr) proteins (<15 kDa) in cancer cell secretomes. Here, we used tricine-SDS-gel-assisted fractionation and LC-MS/MS to systemically identify LMr proteins in the secretomes of five oral cavity squamous cell carcinoma (OSCC) cell lines. Cross-matching of these results with nine OSCC tissue transcriptome datasets allowed us to identify 33 LMr genes/proteins that were highly upregulated in OSCC tissues and secreted/released from OSCC cells. Immunohistochemistry and quantitative real-time PCR were used to verify the overexpression of two candidates, HMGA2 and MIF, in OSCC tissues. The overexpressions of both proteins were associated with cervical metastasis, perineural invasion, deeper tumor invasion, higher overall stage, and a poorer prognosis for post-treatment survival. Functional assays further revealed that both proteins promoted the migration and invasion of OSCC cell lines in vitro. Collectively, our data indicate that the tricine-SDS-gel/LC-MS/MS approach can be used to efficiently identify LMr proteins from OSCC cell secretomes, and suggest that HMGA2 and MIF could be potential tissue biomarkers for OSCC.