The invention provides a mass spectrum model for detecting characteristic proteins of a lung cancer. The mass spectrum model comprises peaks with mass charge ratios of 4054 m/z, 2273 m/z, 4091 m/z, 3955 m/z, 3192 m/z, 3884 m/z and 5337 m/z, wherein in the model, proteins expressing the mass charge ratios of 4054 m/z, 2273 m/z, 4091 m/z, 3955 m/z, 3192 m/z, 3884 m/z and 5337 m/z are serum characteristic proteins of the lung cancer; and when the peaks with the mass charge ratios of 4054 m/z, 2273 m/z, 4091 m/z, 3955 m/z and 3192 m/z are expressed in a down regulation manner and the peaks with the mass charge ratios of 3884 m/z and 5337 m/z are expressed in an up regulation manner, a patient is indicated to be a lung cancer patient or a potential patient. The mass spectrum model can be used for early detection and screening for the lung cancer; the method is simple, easy to operate and high in accuracy; and by utilizing the mass spectrum model, a novel idea is provided for early detection and screening for the lung cancer.