The invention discloses a 
sepsis early prediction method based on 
machine learning. The method comprises the following steps: firstly, extracting clinical data, including 
demographic statistics (suchas 
age and gender), vital sign variables (such as 
heart rate and systolic pressure) and laboratory measurement indexes (such as 
creatinine and 
platelet count), of a patient within 24 hours after a patient enters an ICU by utilizing an 
electronic medical record, preprocessing the data, inputting the preprocessed data into an improved deep forest 
algorithm model for training, and outputting the 
disease probability of the patient after training optimization. And meanwhile, the 
algorithm model can also sort characteristic variables and output an early warning factor which has great influence on 
early prediction of 
sepsis. Finally, corresponding variables of the patient needing to be predicted are input into the trained model, so that 
early prediction of 
sepsis can be carried out on the patient. According to the invention, early prediction is carried out on sepsis based on a 
machine learning method, doctors can be assisted in making clinical decisions, and the prediction accuracy is improved.