Abstract Context Plasma C-reactive protein (CRP) levels are elevated in patients with advanced cancer. Objectives To investigate CRP as a prognostic marker in palliative settings. Methods This multicenter prospective cohort study comprised 2426 patients. Laboratory data were obtained at baseline, and all patients were followed until death or six months after their enrollment. A total of 1511 patients were eligible for the analyses. They were divided into four groups: low-CRP (CRP < 1 mg/dL), moderate-CRP (1 ≤ CRP <5 mg/dL), high-CRP (5 ≤ CRP <10 mg/dL), and very high-CRP (10 mg/dL ≤ CRP) groups. Survival was investigated by the Kaplan-Meier method with the log-rank test. The 30-, 60-, and 90-day mortality rates were tested by Chi-squared tests. Univariate- and multivariate-adjusted hazard ratios (HRs) and 95% CIs in each group were calculated using Cox proportional hazard models. Results Survival rate decreased and mortality rate increased with increasing CRP level. The differences in survival and 30-, 60-, and 90-day mortality rates among the groups were statistically significant ( P  < 0.001). Baseline CRP level was significantly associated with a higher risk of mortality after adjustment for age, gender, primary tumor site, metastasis, chemotherapy, Eastern Cooperative Oncology Group Performance Status, and setting of care (moderate-CRP: HR 1.47 95% CI 1.24–1.73, high-CRP: HR 2.09 95% CI 1.74–2.50, and very high-CRP: HR 2.55 95% CI 2.13–3.05 vs. low-CRP). Conclusion Clear dose-effect relationships between elevated CRP levels and prognoses indicate that CRP could be useful in predicting prognoses in patients with advanced cancer.