The power Internet of Things is an important research field of power grids. Compared to ordinary IoT, it has network points with great density and more equipment types, so higher requirements for network management and security protection are needed. In order to facilitate the management of power IoT terminal equipment, it is necessary to fulfill the adaptation of the access equipment protocol. In response to the issue of diverse protocols among different access device manufacturers, this paper analyzes the traffic data of different power IoT terminal communication protocols and constructs a traffic data preprocessing scheme. By parsing the underlying and application layer protocols, effective application layer payload data is extracted. Furthermore, based on the extracted data and preprocessed features, a power IoT terminal protocol classification method based on the Density-Based Spatial Clustering of Applications with Noise(DBSCAN) is proposed. The model obtained from this method can be deployed on the edge or cloud side of power grid, supplying a foundation for the identification of massive power IoT terminal devices as well as their access protocols in the future. The experimental results show that the accuracy of this method is more than 85%.