The ability to precisely locate sensor nodes has enabled specialized in-network data processing methods within IoT-enabled wireless sensor networks, known as spatial query processing. These queries collect data from nodes situated within user-defined “regions of interest.” Conventional spatial query processing approaches frequently encounter challenges including excessive energy consumption, diminished accuracy, and prolonged processing times. This research focuses on window queries, one of the most prevalent spatial queries used to extract data from nodes within specific two-dimensional regions. We present a strategy for processing window queries in IoT networks that optimizes energy efficiency, response time, and accuracy simultaneously. Our approach introduces intelligent agents that incorporate machine learning, knowledge representation, and autonomous decision-making capabilities specifically tailored for spatial query processing across geographical zones. Experimental evaluations demonstrate significant improvements: 35% reduction in energy consumption, 99.1% query accuracy (compared to 96.8% for baseline methods), and up to 71% decrease in latency under various network densities and query area sizes.