Fishing activities result in a depletion of fish resources and an increase in bycatch probability of non-target species, which are thought to be the cause of the recent demographic change experienced by coastal cetaceans in waters off Guangdong province, China. However, fishing intensity and its spatiotemporal pattern remain poorly examined in this region. Here, random daily tracks from 100 fishing vessels were used to estimate fishing operations in the coastal waters off Guangdong province. The performance of two methods was tested with Automatic Identification System (AIS) data collected in the Pearl River Delta region (PRD). This assessment found that the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) could falsely identify port regions as fishing grounds, while the speed-based method provided a more accurate estimate of fishing grounds with less computational effort. Subsequently, the spatiotemporal pattern of fishing intensity in the coastal waters off the entire Guangdong province was examined with the speed-based method. Our results showed that 87.75% of the coastal fishing activities occurred in waters <40 m in depth. The method detected two spatially heterogeneous fishing grounds extending from the PRD to either the west or east sides of Guangdong province. In the west fishing ground, 43.4% of fishing vessels operated in the waters off Yangjiang city; while in the east fishing ground, fishing activities were more evenly distributed. The present study suggests that the AIS is able to provide high-resolution data to monitor the development of fishing activities, which may be useful for informing marine resource management and spatial planning.