With the rapid development of global maritime trade, the demand for visual analysis of ship activities is growing. Without very professional knowledge, users are difficult to quickly select appropriate visual methods to analyze ship activities, and face the problem of "cold start". Based on multisource heterogeneous ship activity data, this paper constructs a ship activity knowledge graph including four types of ship activity elements of "time, place, ship and activity". Combining with the needs of ship activity analysis, four analysis tasks are designed, including ship object feature analysis, ship activity event analysis, ship activity correlation analysis and ship activity space-time analysis; Besides, based on the knowledge graph of ship activities, the visual analysis method of ship activities based on activity elements is designed, and the visual analysis model of ship activity knowledge graph is designed combined with the interactive modes of filtering, association, and navigation. In this paper, the AIS data and the "boiling ring trade" maritime event data are used as the ship activity data. Based on the above model, a prototype system for visual analysis of the ship activity knowledge graph is developed. The results show that the system can recommend relatively appropriate visual methods for users, provide more perspectives for the analysis of ship activities, and also verify the conclusions such as the changes in the laws of ship activities caused by special circumstances such as the closed fishing periods.