Objective To study the value of apparent diffusion coefficient (ADC) difference value between cerebral infarction (CI) in normal and affected area in evaluating the risk of hemorrhagic transformation (HT). Methods 50 cases of patients with early CI who were treated in the department of neurology of our hospital between July 2013and June 2015 were included in the study. All patients were given spiral CT and MRI routine brain scan after admission and all of them underwent DWI and ADC scan. On MR ADC images, ADC difference value between CI normal and affected area was measured and calculated; 2 weeks after attack, the patients were followed up and underwent MR reexamination. The occurrence of HT was observed. The correlation between HT and ADC difference value was further analyzed. Results Through follow-up and reexamination, among 50 cases of patients with CI, 5 cases were with HT and the incidence of HT (10%) was lower than that of non-HT (NHT) (90%). Early ADC difference measurements showed that there were 5 patients with >205.5×10-6mm2 /s and 5 patients with HT were in complete agreement with follow-up and reexamination. ADC difference prediction showed better accuracy and specificity. ADC difference of patients with HT [(221.4±10.2)×10-6 mm2 /s] was higher than NHT [(145.3±11.4)×10-6mm2 /s] (P<0.05). Conclusion ADC difference measurement between early CI normal and affected area can better reflect hemorrhage in infarcted lesions. This study shows that ADC difference >205.5×10-6mm2 /s prompting high risk of HT, which is of high reference value in clinical evaluation and prediction of HT and guiding treatment.