Objective To explore a simpler, more economic and reproducible method to reproduce a model of high oxygen induced acute lung injury (HALI) in rats. Methods An animal feeding box equipped with a controllable high oxygen was designed. 100 Sprague-Dawley (SD) rats were divided into normal control group and HALI group by random number table method, with 50 rats in each group. Each group was randomly subdivided into five subgroups according to the duration of exposure to high oxygen, namely 0, 24, 48, 72 and 96-hour subgroups, with 10 rats in each subgroup. The rats in normal control group were kept in cages with ambient air, and the rats in HALI group were kept in an oxygen tank in which the oxygen concentration was higher than 90% volume ratio, with the temperature maintained at 25-27 ℃, humidity of 50%-70%, and CO2 concentration < 0.5% for 23.5 hours every day. The arterial blood of rats was collected for analysis of blood gas at all time points, and the oxygenation index (OI) and respiratory index (RI) were calculated. Then the rats were sacrificed and the right lung was harvested, which was sectioned and stained with hematoxylin and eosin (HE). The changes in histopathology were observed with light microscopy, and pathological score was recorded. The left lung was harvested for the measurement of the wet/dry weight ratio (W/D). Results With the prolongation of high oxygen exposure time, the degree of lung injury in HALI group was gradually increased, and the degree of derangement of alveolar structure appeared in an increasing degree, with destruction of the alveolar wall, widening of alveolar space, and appearance of edema, and inflammatory cell infiltration. A small quantity of red blood cells exudation could be found in some rats. The pathologic changes were most obvious at 48-72 hours after exposure. With the prolongation of high oxygen exposure time (0, 24, 48, 72, 96 hours), the OI (mmHg, 1 mmHg = 0.133 kPa) in HALI group was gradually decreased (446.67±29.93, 306.19±37.23, 269.70±29.00, 253.81±43.40 and 245.58±35.25), RI, pathological score of lung tissue and W/D ratio were gradually increased [RI: 0.25±0.04, 0.31±0.06, 0.38±0.06, 0.46±0.07 and 0.44±0.03; pathological score of lung tissue: 0.00±0.00, 0.90±0.74, 2.90±1.20, 4.70±1.57 and 4.80±1.23; lung W/D ratio: 3.84±0.61, 4.14±0.46, 4.56±0.34, 5.32±0.27 and 5.18±0.25]. Statistically significant differences were found in 72-hour group as compared with that of other groups (all P < 0.05), while no significant difference was found between 96 hours and 72 hours groups (all P > 0.05). There were significant differences in changes between 24, 48, 72, and 96 hours as compared with those of the normal control group: OI (mmHg): 24 h 306.19±37.23 vs. 435.65±25.34 and 96 h 245.58±35.25 vs. 465.42±24.75; RI: 24 h 0.31±0.06 vs. 0.24±0.04 and 96 h 0.44±0.03 vs. 0.24±0.06. The same as true in pathological scores of lung tissue: 24 h 0.90±0.74 vs. 0.00±0.00 and 96 h 4.80±1.23 vs. 0.00±0.00; lung W/D ratio: 24 h 4.14±0.46 vs. 3.79±0.44 and 96 h 5.18±0.25 vs. 4.12±0.91, all P < 0.05. Conclusions A self-designed high oxygen box is simple, easy to operate and reproduction of HALI model can be attained. Sustained exposure to high concentrations of oxygen (≥ 90%) for 24 hours can replicate the HALI model successfully, and the most serious injury appears at 48-72 hours after exposure.