The Automatic Identification System (AIS) device is mandatory for ships that comply with the International Convention for the Safety of Life at Sea (SOLAS). AIS is intended for vessel traffic monitoring to improve shipping safety. In the examined area, the base station received 22 128 345 messages in April 2019. Approximately 80% of these messages included position reports, which were subjected to geospatial analysis. One possible utilization of AIS messages is used in an intelligent maritime transport statistics production system called TranStat in the Gospostrateg project. This specific study compares the speed of executing geospatial queries in a relational PostgreSQL database engine and a non-relational MongoDB database engine. For the purpose of this research, we have defined four AIS datasets, four test polygons of varied number of vertices, and a reference point on a fairway. The tests were used to assess the execution of the queries in a database that returns the number of ships located in a predefined area and the number of ships located at a preset distance from the defined point. It has been determined from the test results that test queries are performed faster and data stored in the database occupy less disk space in MongoDB than in PostgreSQL. Faster geospatial analysis of AIS messages may improve the navigation safety by earlier detection of dangerous situations.