Aim Construct a nomogram for predicting hospital mortality in critically ill patients with heart failure (HF) from Medical Information Mart for Intensive CareⅢ (MIMIC-Ⅲ) database. Methods Data were extracted involving critically ill patients with HF from MIMIC-Ⅲ database. All eligible patients (n=8 604) were randomly classified into the training set (n=6 022) and validation set (n=2 582) with a ratio of 7∶3, and the outcome was hospital mortality. LASSO-Logistic analysis in the training set was used to determine the prognostic factors and the nomogram for predicting hospital mortality was thereby constructed. Receiver operating characteristic (ROC) curve, calibration curve, decision analysis curve (DCA) and clinical impact curve (CIC) were generated to assess the discrimination, calibration and clinical utility of the nomogram, respectively. Results LASSO-Logistic analysis showed that red blood cell distribution width (RDW), respiratory rate, oxygen saturation, acute physiology score Ⅲ (APSⅢ) and simplified acute physiology score Ⅱ (SAPSⅡ) were independent predictors. In both training and validation set, the area under the curve (AUC) was 0.775 (95%CI:0.757～0.792) and 0.767 (95%CI:0.742～0.793), respectively. Calibration curve was highly consistent with the diagonal line, and the mean absolute error was 0.009 and 0.016. AUC and calibration curve showed great performance of the predictive model in discrimination and calibration. Meanwhile, DCA and CIC revealed that the predictive model provided significant net benefits for most threshold probabilities. Conclusion Nomogram is a simple and accurate tool for predicting hospital mortality in critically ill patients with HF.