The XML-enabled Spatial Visualizing Database can be widely used because of its high security and convenience of management. It is often implemented by relational databases. But, these databases can't index points, curves and surfaces separately based on their characteristics. In addition, the indexing systems of these databases are not capable of creating the function of indexing on a wide range of relational databases. These limitations were solved by the following two steps in the paper. Firstly, referring to the SVG spatial expressional model, bring up a solution to index points, curves, and surfaces separately with different algorithms. Secondly, elaborately design a spatial indexing system with high cohesion and low coupling by using three design patterns: Template Method, Bridge and Adapter. Eventually, our system has three achievements: 1) separately index points, curves and surfaces with different algorithms catering for the users; 2) index SVG visualizing data on heterogeneous relational databases; 3) many of the open-source spatial indexing projects can be used in this system after modifying some codes. Experiments have been carried out to prove the
conclusions mentioned above.