Objective To explore the predictive value and significance of hemorrhage after thrombolysis （HAT） model combined with peripheral blood neutrophil to lymphocyte ratio （NLR） in the hemorrhage transformation （HT） of acute cerebral infarction （ACI）. Methods Clinical data of 433 patients with ACI who were admitted to the Affiliated Hospital of Xuzhou Medical University from May 2018 to April 2019 were collected and analyzed retrospectively. On the basis of hemorrhage transformation during hospitalization， all the participants were divided into two groups， non-hemorrhagic transformation group （NHT group） and hemorrhagic transformation group （HT group）. The general clinical data of the two groups were collected and analyzed by single factor analysis， including demographic basic data （age， gender）， vascular risk factors （smoking， drinking， hypertension， diabetes， atrial fibrillation）， blood pressure （baseline systolic pressure， baseline diastolic pressure）， baseline NIHSS score， trail of Org 10172 in acute stroke treatment trial （TOAST） typing and laboratory test results， such as leukocyte count， neutrophil count， lymphocyte count， NLR， baseline blood glucose and CT early infarct area. The NLR， HAT model score and other clinical data were compared between the two groups. Multiple logistic stepwise regression analysis was used to compare the related factors of hemorrhage transformation in patients with ACI who were not treated with thrombolysis， and ROC curve was used to analyze the predictive effect of HAT model （including three factors of patients' previous diabetes history or baseline blood glucose， CT early infarct area and NIHSS score） combined with peripheral blood NLR on hemorrhage transformation in patients with ACI who were not treated with thrombolysis. Results A total of 433 patients were included， among that 77 patients in the HT group， counted for 17.8%. The other 356 patients were in the NHT group. The baseline systolic blood pressure， history of atrial fibrillation， leukocyte count， neutrophil count， NLR， baseline blood glucose and NIHSS scores of HT group were higher than those of NHT group， while the lymphocyte count was lower than that of NHT group （P ＜ 0.01）. There were significant differences in the early infarct area， TOAST classification and HAT model score between the two groups （P ＜ 0.01）， and there was no significant difference in the other factors （P＞ 0.05）. Multivariate logistic regression analysis showed that NLR （OR=1.23， 95%CI=1.12-1.34）， baseline blood glucose （OR=1.18， 95%CI=1.04-1.33）， baseline national institute of health stroke scale （NIHSS） score （OR=1.04， 95%CI=1.00-1.08）， baseline systolic blood pressure （OR=1.02， 95%CI=1.01-1.04）， early infarct volume of CT （OR=2.41， 95%CI=1.38-4.22）， and cardiogenic stroke were all independent risk factors of HT （P＜ 0.05）. ROC curve analysis showed the optimal cut-off value of HAT model score combined with NLR was 1.50， the sensitivity and specificity were 83.10% and 62.40% respectively， and the area under ROC curve was 0.81 （95%CI=0.75-0.86，P ＜ 0.05）. Conclusions The greater the NLR， baseline blood glucose， baseline NIHSS and early infarct area in CT， the greater the risk of HT. The HAT model score combined with NLR can be used as a predictor of HT in ACI patients without Intravenous thrombolytic therapy.