This study is aimed at investigating the prognostic biomarkers of patients with stage IIIA-N2 non-small-cell lung cancer (NSCLC) and at analyzing the correlation between tumor mutation load (the frequency and number of tumor mutations) and prognosis. Clinical data of 35 patients with stage IIIA-N2 NSCLC were collected from Cancer Hospital, Chinese Academy of Medical Sciences. Whole blood samples from the peripheral vein were taken at different treatment periods, and the mutations of cell-free DNA (cfDNA) were detected. Multivariate analysis showed that smoking (P = 0.0308), mutation number > 2 (P = 0.0283), and max mutation frequency > 0.025 (P = 0.0450) were associated with improved progression-free survival (PFS). The overall survival (OS) of well-differentiated NSCLC patients was better than that of poorly differentiated ones (P = 0.0006). The rates of PFS, disease-free survival, local-regional recurrence-free survival, and local-regional progression-free survival were significantly higher in the group with a mutation number > 2 than in the group with a mutation number ≤ 2. The mutation number of the preoperation group was significantly higher than that of the postradiochemotherapy group (5 vs. 2.5, P = 0.023), and the max mutation frequency change was approximately significant in the postradiochemotherapy group compared with the postoperation group (2.6% vs. 1.85%, P = 0.067). The max mutation frequency is positively correlated with vascular invasion (21.13% vs. 3.62%, P = 0.04). Furthermore, Met, ALK, APC, PTEN, ERBB4, NF1, and other genes, involving multiple tumor suppressor genes and lung cancer-driven genes, did not mutate in recurrence-free patients when compared with recurrent patients. In conclusion, differentiation, smoking, mutation frequency > 0.025, and mutation number > 2 are prognostic factors. The frequency and number of gene mutations in cfDNA are expected to be prognostic predictors of NSCLC.