Objective To explore the stability of rat models of subdural hematoma prepared by subdural injection of different volumes of autologous blood. Methods The rats were randomly divided into sham group (36), 300 μL blood group, 500 μL blood group, and 700 μL blood group (each group 60 rats). The rats of model groups received subdural injection of 300 μL, 500 μL, or 700 μL autologous blood, respectively. At the postoperative 2nd, 4th, 6th, 8th, 10th, 14th days, blood samples were taken from the abnormal aorta, and the brains were taken out for gross examination and taking photographs, six rats were used for each time. Enzyme linked immunosorbent assay (ELISA) was performed to determine the content of serum NSE and S100B proteins in the rats in each group. Results Compared with the sham operation group, the serum NSE in the 300 μL group was significantly increased at the 2nd and 4th days (P<0.01), and then gradually reduced at the 6th and 8th days (P<0.05), indicating that the hematoma began to disappear, and at the 10th and 14th days returned to a similar level of the sham operation group (P>0.05). In the 500 μL and 700 μL blood groups, the NSE contents at 2nd, 6th, 8th, 10^th and 14th days were significantly increased (P<0.01), but not significantly increased at the 4th day (P>0.05). The content of S100B protein in the 300 μL blood group was significantly higher at the fourth day (P<0.01), lower at the 2nd and 6th days (P<0.05), and at 8th, 10th and 14th days was similar to that in the sham operation group (P>0.05 for all), indicating that the hematoma disappeared gradually, and the damages repaired. The S100B protein content of the 500 μL and 700 μL blood groups was constantly kept at a higher level (P<0.05). Conclusions Compared with the 300 μL ad 700 μL blood groups, the rat model of subdural hematoma developed by subdural injection of 500 μL autologous blood is the best, and can be used for studies of rat subdural hematoma.