Accurately positioning wireless sensor network nodes is challenging when data is scarce. Traditional methods heavily rely on data-driven or content-specific approaches, making precise localization difficult. This paper introduces a novel method combining human intelligence and machine learning to address this issue. By integrating the Three-Dimensional (3D) Voronoi diagram division and a hybrid regression model, it aims to enhance localization accuracy with limited data. The proposed approach involves two stages: offline training and online testing. During the offline phase, Voronoi diagram division segments the localization space into smaller regions, reducing the need for human intervention. The hybrid model, called Hybrid Weighted Regression with Support vector regression and K-nearest neighbors Regression (HWR-SKR), combines Support Vector Regression (SVR) and K-nearest Neighbors Regression (KNR) to leverage the strengths of both. Training and testing utilize received signal strength data, anchor node coordinates, and Voronoi cell vertex coordinates. Experiments demonstrate that the proposed method accurately locates nodes even with limited data. The HWR-SKR model outperforms individual SVR and KNR models, improving real-time positioning tasks’ accuracy and stability. This study presents a promising solution for precise node localization in wireless sensor networks, enhancing localization performance and supporting sensor network applications requiring accurate spatial awareness.