This experiment contains eight datasets: Aru Data, Naimona’nyi Data, Guliya Data, Parlung Data, Dunde Data, Dagze Data, XiaodongKemadi Data and Southeastern data (China southeast of Qinghai-Tibet plateau Data). The Spatial Modeling of Near-surface Air Temperature in the Glacierized Areas of the Tibetan Plateau are generated by integrating observation temperature data from glaciers in the high-altitude areas of the Qinghai-Tibet Plateau and thermal infrared remote sensing data. Temperature spatialization in glacierized regions involves converting discrete station observations into continuous, physically consistent gridded fields, serving as crucial input data for studying climate-glacier interactions. In such cases, spatialized air temperature is derived exclusively from ground-station measurements, interpolated within a geographic framework constrained by Landsat 8 thermal infrared data. The remote sensing data solely provide spatial coverage (30 m resolution), serving as a geometric reference to ensure alignment accuracy. The L2SP image data are atmospherically corrected and the Digital Number (DN) of the standard L2SP is stored in a 16-bit unsigned integer format. Single-point outliers or missing values were filled using a predefined mask or corrected via Kriging interpolation. The interpolation procedure allocated stations to Voronoi cell centroids, with nearest neighbors identified via brute-force search. Delaunay triangulation was then constructed using both station coordinates and Voronoi cell vertices, enabling trigonometric interpolation of temperature fields. The initial near-surface air temperature distribution was subsequently filtered and smoothed to mitigate noise and enhance spatial coherence. This dataset supports climate research, glacier melt modeling, and water resource assessments, along with other applications requiring reliable temperature spatialization in glacierized regions.