Textual location descriptions provide meaningful representations of spatial positions and are widely used to convey the locations of people and objects. Contemporary approaches for generating location descriptions focus mainly on point-based queries, such as reverse geocoding and human-based generators. In this study, an ontology-based textual location description framework that smoothly integrates spatial data, spatial cognition, and locational semantics is proposed. The aim is to generate automated natural-language location descriptions for multidimensional feature queries. Transportation scenarios are used as experimental cases. This method outperforms human-written and generative artificial intelligent (AI) descriptions by more than 30%. The proposed framework provides a new path for location description, supporting diverse spatial queries and contextual and spatial-cognitive applications.