Background and purpose:How to predict the survival length for terminally cancer patients is very important, it will help families and physicians to make decisions. This study aimed to reveal the factors related to the survival time of terminally ill cancer patients who received palliative care in our hospital. Methods:The clinical data of 271 dead patients treated in the Department of Palliative Care, Fudan University Shanghai Cancer Center from Mar. 2007 to Mar. 2012 were analyzed. The Kaplan-Meier method and the log-rank test were used to determine the corresponding factors with survival. Cox regression model was used to examine the independent prognostic factors. Different survival length of groups divided by different prognostic indexes was compared by log-rank test. Results:Seven factors were found to be related with the survival according to univariate analysis. The related factors were Karnofsky performance score (P<0.001), dyspnea (P=0.037), delirium (P=0.015), high white blood cell count (P=0.012), low lymphocyte percentage (P=0.030), high lactate dehydrogenase (P<0.001) and low serum albumin (P=0.001). The multivariate analysis selected four independent factors:Karnofsky performance score<30, high lactate dehydrogenase, low serum albumin and delirium.Conclusion:The study shows the clinical survival prognostics with Chinese characteristics. The combination of the seven factors may be useful but more studies in this area deserve further investigated.