In view of such flaws as effort demanding and time consuming disadvantages,along with missed diagnosis and misdiagnosis to be found in pathological analysis due to the difference of doctors’experience,an automatic classification method of non-small cell lung cancer pathological image,with the method of multi-feature fusion combined,has thus been proposed based on high-efficiency network (EfficientNet) model. Firstly,EfficientNet model is used to train the preprocessed pathological image,with the multi-dimensional depth features to be extracted. Then,the extracted GLCM features is fed into the CNN for a further extraction of the texture features of the image. Finally,all the fused features are to be fed into a Softmax classifier,thus establishing a NSCLC fusion diagnosis model based on pathological image classification. Theoretical analysis and experiments show that the area under the receiver operating characteristic (ROC) curve (AUC) of this method reaches as high as 86%,verifying the fact that the accuracy is also improved compared with other classification methods.