Background: Adjuvant chemotherapy (ACT) has been recommended for stage III colon cancer patients in NCCN guidelines. However, the benefit of ACT is still controversial in elderly patients. Methods: We retrospectively included stage III colon cancer patients with aged over 70 that were operated between 2008 and 2015 in Tosei general hospital from Japan. We calculated 3 years recurrence rate and performed a multivariate analysis to estimate the risk of recurrence. The model was adjusted by including the following 7 confounders: Age, ECOG-PS (0-1 or ≥ 2), tumor site (right or left), T stage (T1-3 or T4 ), N stage (N1 or N2), pathological type (por-sig or other), and complete ACT (yes or no). No complete ACT was defined as discontinuation ACT and surgery alone. Results: 182 patients were identified (87 patients aged < 70 and 95 patients aged ≥ 70). ACT was performed in 58 patients (61.1%) and complete rate was 67.2% in elderly patients. The median follow up was 36.0 months. 29 patients (30.0%) had relapse. 3 years recurrence rate was 24.3% and 37.8% with complete ACT or no complete ACT (p = 0.29). In multivariate analysis, N2 was only the risk of recurrence (HR 6.95, p < 0.01). In further analysis, addition of oxaliplatin was significant risk factor of recurrence in elderly patients (HR 10.4, p < 0.01). On the other hand, in young patients, no complete ACT was only risk of recurrence (p = 0.01). 3 years survival rate was 89.1% and 81.0% with complete ACT or no complete ACT (p = 0.27). Conclusions: The multivariate analysis showed that ACT did not reduce the risk of recurrence in elderly patients. Moreover, addition of oxaliplatin suggested not only futile but harmful in those patients. These results support the previous reports (MOSAIC study and PETACC-8 trial) that microsatellite instability and RAS status may affect the effect of oxaliplatin.