The invention discloses a kit for predicting the lung injury clinical progress of a patient with chest tumor after radiotherapy. A detection module detects the number and proportion of seven immune cells in peripheral blood of the patient with chest tumor, and the seven immune cells comprise NK cells, lymphocytes, granulocytes, monocytes, DC cells, T lymphocytes and B lymphocytes; a measuring module is used for measuring the concentrations of five cytokines in peripheral blood, and the five cytokines comprise interleukin 4, interleukin 6, interleukin 8, interleukin 13 and gamma interferon; anacquisition module acquires the clinical information of the patient withchest tumor, and the clinical information comprises age, gender, weight, hypertension, diabetes, pulmonary infection, radiationfield and radiation dose; and a prediction module inputs the number and proportion of the seven immune cells, the concentrations of the five cell factors and the clinical information into a feedforward neural network for prediction, and predicts the probability of occurrence and severity of lung injury at different stages after radiotherapy of the patient with chest tumor .