Location analysis and geospatial data mining are significant approaches to analyzing huge datasets. Geospatial artificial intelligence (GeoAI) uses deep learning and artificial intelligence to evaluate geographic data. This might have far-reaching implications for a variety of sectors, including but not limited to urban planning, environmental monitoring, and disaster response. Complex mining processes will be necessary when the number of geographical data resources grows dramatically. This paper presents a Geographic Deep Mining Network (GDMN), which uses deep learning to analyze geographical data more efficiently and precisely. The major objective for establishing the network was to support geospatial applications. The establishment of the GDMN aimed to enhance the efficiency of geographical data processing. This applies especially to AI today. This paper introduces the geographic deep mining network (GDMN), a deep learning-based spatial AI technique. The proposed technique includes the Geospatial Data Fusion Module (GDFM), Spatial Feature Extractor (SFE), and Temporal Geospatial Attention Mechanism. An in-depth analysis compares GDMN to six well-known algorithms and demonstrates improvement in several areas. These measurements include recall, accuracy, precision, F1-score, IoU, training time, and inference time. We observed that GDMN improves map AI in data mining and location analysis. Geospatial intelligence is always evolving and finding new applications in urban planning and environmental tracking. This study contributes to it. The construction of a sturdy, versatile structure does this.