Introduction The Patient-Generated Subjective Global Assessment (PG-SGA) is currently the standard nutritional assessment tool for patients with cancer. In a retrospective assessment of a prospective cohort, we showed that the Nutritional Risk Index (NRI) seemed to be associated with treatment toxicity and survival in patients with metastatic colorectal cancer (mCRC). Objective The objective of this study was to compare these two nutritional tools (PG-SGA and NRI) on their correlation with chemotherapy-related toxicity and survival in non–pre-treated patients with mCRC. Methods This prospective multicentre observational study enrolled non–pre-treated patients with mCRC. PG-SGA and NRI were performed at the onset of first-line chemotherapy. Treatment-related toxicities were registered according to National Cancer Institute Common Toxicity Criteria Adverse Event version 4.0. Progression-free survival (PFS) and overall survival (OS) were calculated from the start of treatment. Results A total of 168 patients were included from eight French centres. Patients were considered malnourished in 41% of cases according to PG-SGA and 56% of cases according to the NRI. In multivariate analysis, malnutrition according to PG-SGA was significantly associated with chemotherapy-related grade ≥2 clinical toxicities (odds ratio: 3.7; 95% confidence interval [CI]: 1.7–8.4; p = 0.001) and OS (hazard ratio [HR]: 2.6; 95% CI: 1.3–5.3; p = 0.006), but not with PFS (HR: 1.5; 95% CI: 0.8–2.6; p = 0.2). Conversely, malnutrition according to the NRI was not significantly associated with these tolerance and efficacy parameters. Conclusion Although more complex to perform in daily oncology practice, the PG-SGA score appears to be the best nutritional assessment tool because of its strong association with clinically relevant oncological outcomes such as OS and treatment-related toxicities in patients with mCRC.