Background: Radiation-induced lung injury is a refractory side effect in lung cancer radiotherapy, the mechanism still remains unclear, hence an appropriate animal mode may become useful to investigate it. Materials and Methods: 50 female Wistar rats were randomly divided into 5 groups, average 10 rats/cage: A. control group B. 3Gy×10f C. 6Gy×5f D. 12.5Gy×1f E.15.3Gy×1f. By different doses of radiation to the right lung of rats under American Varian linear accelerator, all the rats were executed at 30days after irradiation, lung tissues were collected to observe the pathological changes by HE staining , and additional lung tissues were extracted mRNA to measure gene expression of TNF-a and IL-18 by qRT-PCR. Results: In contrast to the control group, congestion, gore, edema, broadening alveolar interval and the bulla can be observed after irradiation. 3Gy×10f radiation led to most rat death because of frequent anaesthesia. 12.5Gy×1f radiation induced mild lung injury. 6Gy×5f and 15.3Gy×1f radiation induced remarkable lung injury. However, 15.3Gy×1f was seldom adopted in clinical settings. The measurement of cytokine variation by qRT-PCR also indicated that 6Gy×5f radiation stimulated the most obvious elevation of IL-18, while 15.3Gy×1f radiation elevated TNF-a most. Conclusion: A rat model of radiation-induced lung injury was appropriately established, 6Gy×5f radiation was the most suitable radiation method for establishing this rat model.