With the continuous increase of data in the power Internet of Things, the existing power Internet of Things based on cloud computing architecture cannot meet the industry’s ever-growing data processing needs. Based on the edge computing architecture, this paper designs a data classification and processing method for the power Internet of Things based on the cloud-edge collaborative architecture. This method summarizes the collected raw data at the edge layer, uses the random forest algorithm to classify and process the data that needs to be processed in real time, and feeds back the processing results to the device; At the same time, the edge layer uploads the data that needs to be processed in the cloud, and uses the LSTM-FCN data classification model to classify and store these data at the cloud layer. Finally, the accuracy of this method is verified by experiments.