Text2SQL has evolved into a significant tool for data analysis and database management by reducing the technical barriers to non-expert users interacting with relational databases. The introduction of large language models, represented by GPT, further improves the performance of Text2SQL systems. However, the existing Text2SQL techniques are difficult to apply directly to the spatial database domain because spatial data involves complex geometric relationships, diverse query types, and the need for high-precision semantic understanding. A natural language query transformation method for spatial databases is proposed to solve the above problems and reduce the threshold of interaction between non-experts and spatial databases. The method consists of two core phases: (i) natural language understanding and (ii) structured language generation. In phase (i), an entity information extraction algorithm is used to extract key entities, and a spatial data query corpus is constructed based on large language models to determine the query type. In phase (ii), the structured language model is selected according to the query type, and then the entities are mapped into the structured language model to generate executable database queries. Experimental results on several real-world datasets show that the proposed method achieves efficient translation from natural language queries to executable languages of spatial databases.