The purpose of this study is to develop and validate a scale prognostic of survival in hospitalized, terminally ill cancer patients in China. Terminally ill cancer patients hospitalized at two general hospitals in China were prospectively analyzed. Patients were divided into a training cohort (n = 181) and a testing cohort (n = 128). Factors prognostic of survival were identified in the training cohort and combined into a scale, which was validated in the testing cohort. In the training cohort, eight factors associated with reduced survival were identified: low performance status, dyspnea at rest, reduced oral intake, cognitive impairment, edema, leukocytosis, and elevated urea and alanine transaminase concentrations. A prognostic prediction score was calculated for each patient, based on the weight of these eight predictors in the regression model, with scores ranging from 0 (no altered variables) to 12 (maximal altered variables). Patients with different prognostic scores had significantly different prognoses (p < 0.001). A cutoff point of a parts per thousand yen4 was optimal in categorizing patients with "low" (score < 4) and "high" (score a parts per thousand yen4) risk of survival for less than 30 days, with median survival time in these groups of 47 and 9 days, respectively. Using this cutoff point on the testing cohort, median survival time for the low and high risk groups were 66 and 11 days, respectively. We identified eight indicators predictive of poor survival in Chinese patients hospitalized with terminal cancer. A prognostic scale that includes these indicators may help in making decisions about end-of-life care.