We aimed to assess the prognostic value of the pre-operative GRADE score for long-term survival among older adults undergoing major surgery for digestive or non-breast gynaecological cancers. Between 2013 and 2019, 136 consecutive older adults with cancer were prospectively recruited from the PF-EC cohort study before major cancer surgery and underwent a geriatric assessment. The GRADE score includes weight loss, gait speed at the threshold of 0.8 m/s, cancer site and cancer extension. The primary outcome was post-operative 5-year mortality. Patients were classified as low risk (GRADE ≤ 8) or high risk (GRADE > 8) on the basis of the median score. A Cox multivariate proportional hazards regression model was performed to assess the association between pre-operative factors and 5-year mortality expressed by adjusted hazard ratio (aHR) and 95% CI. The median age was 80 years, 52% were men, 73% had colorectal cancer. The 30-day post-operative severe complication rate (Clavien-Dindo ≥ 3) was 37%. The 5-year post-operative mortality rate was 34.5%. A GRADE score ≥ 8 (aHR = 2.64 [1.34–5.21], p = 0.0002) was associated with post-operative mortality after adjustment for Body Mass Index < 21 kg/m2 and Instrumental Activities of Daily Living <3/4. By combining very simple geriatric and cancer parameters, the pre-operative GRADE score provides a discriminant prognosis and could help to choose the most suitable treatment strategy for older cancer patients, avoiding under or over-treatment.