Objectives: Low fat-free mass index (FFMI) has been related to a higher mortality in community populations. However, information on the relationship between FFMI and mortality is lacking for patients with cancer. The objective of this study was to examine the association between FFMI and all-cause mortality in Chinese cancer patients. Methods: This retrospective analysis included data on 1744 patients with cancer from a multicenter cohort study. The restricted cubic splines were used to flexibly model the association of FFMI with all-cause mortality. The association between low FFMI and overall survival was analyzed with the Kaplan-Meier method and a Cox model. Results: Among all patients, there were 702 men (40.3%) and 1042 women (59.7%). The optimal cutoff point of low FFMI was 16.31 for men and 14.14 for women. The FFMI showed an inverse association with all-cause mortality for men (per standard deviation [SD] increment; hazard ratio [HR]: 0.72; 95% confidence interval [CI], 0.60-0.87; P < 0.001) and a nonlinear relation for women (per SD increment; HR: 0.88; 95% CI, 0.78-0.99; P = 0.048). After adjustment, a low FFMI score was independently associated with an increased risk of mortality for both men and women. In addition, FFMI showed a strong L-shape (per SD increment; HR: 0.59; 95% CI, 0.46-0.76; P < 0.001) relation with all-cause mortality in elderly patients with cancer. For specific tumor type, a low FFMI score was independently associated with worse prognosis in patients with lung and upper gastrointestinal cancer. Conclusions: A low FFMI score was associated with all-cause mortality in patients with cancer, especially for elder adults with cancer. These results highlight the usefulness of the FFMI for routine clinical assessment and prognostic estimation in patients with cancer. (c) 2021 Elsevier Inc. All rights reserved.