Abstract: In view of the low spatial resolution and strong susceptibility to solar radiation effects of satellite-based thermal infrared remote sensing in geothermal exploration, this study employs a DJI Mavic 3T unmanned aerial vehicle to acquire high-resolution true-color and thermal infrared imagery over the study area. An automated land surface temperature retrieval workflow was developed based on the DJI Thermal SDK to obtain UAV-derived surface temperature data. Validation using in situ temperature measurements demonstrates that the UAV-retrieved temperatures exhibit a high degree of consistency with the observed temperature variation trends at ground sampling points. After resampling the UAV thermal infrared imagery to 30 m, the retrieved temperatures show strong spatial agreement with those derived from Landsat-8 TIRS data. Based on the proposed “cross-shaped” and “linear” geothermal target delineation patterns, six potential geothermal targets were identified by integrating the regional structural framework with the spatial distribution characteristics of thermal anomalies. Furthermore, the geothermal genesis of the delineated targets was analyzed through the combined use of measured shallow ground temperature data and existing geological information. The results indicate that UAV-based thermal infrared technology enables high-precision identification of geothermal anomalies in complex mountainous terrains, providing effective technical support for small-scale geothermal target delineation and resource potential assessment. Future work should incorporate multi-depth ground temperature measurements and shallow heat diffusion models to strengthen the quantitative linkage between surface thermal anomalies and deep geothermal heat sources.