Objective To investigate the relationship between Oxfordshire community stroke project (OCSP) classification and MRI classification in acute cerebral infarction. Methods A total of 282 patients with acute cerebral infarction were retrospectively evaluated with OCSP classification and imaging characteristics. Results According to OCSP classification, of all 282 patients with acute cerebral infarction, 32 (11.3%) experienced total anterior circulation infarction (TACI), 86 (30.5%) partial anterior circulation infarction (PACI), 111 (39.4%) lacunar infarction (LACI), and 53 (18.8%) posterior circulation infarction (POCI). The consistency was found in 201 cases (71.3%) between the OCSP classification and imaging classification, with the accuracy of 77% (27/35) for TACI, 79% (42/53) for PACI, 69% (95/137) for LACI and 65% (37/57) for POCI. Conclusion OCSP classification can predict the location and size of cerebral infarction with a high accuracy, and is well consistent with the MRI findings.