With the continuously increasing requirement of data processing speed, it is more and more suitable for images to be processed on cloud computing platform with distributed parallel computing. Therefore, this paper studied the rapid classification of massive images based on cloud computing platform. The multi-mode information fusion of massive images was carried out, which gave the implementation steps of the proposed massive image fusion method based on information complementation. This paper proposed to classify the massive images based on salient interactive feedback network. The model constructed in this paper not only simplified the complicated network parameter initialization and tuning process, but also avoided the classification failure caused by the scarcity of labeled images, which is more suitable for application and practical application scenarios. The experimental results verified the effectiveness of the constructed model in rapidly classifying massive images.