Abstract: Objectives: Landslide monitoring requires both extensive spatial coverage and continuous temporal data, which single-technology solutions struggle to provide. While GNSS delivers precise continuous point measurements, its sparse spatial sampling cannot fully characterize fragmented landslide areas. UAV photogrammetry offers dense spatial data, but its temporal resolution and accuracy remain inadequate for continuous monitoring. This study develops a data fusion method integrating both technologies to achieve continuous, centimeter-level areal monitoring of landslide deformation. Methods: The method involves repeated UAV surveys over a landslide area equipped with GNSS monitoring stations. UAV trajectory data are processed using post-processed kinematic (PPK) positioning and cubic spline interpolation, establishing a unified coordinate framework with the GNSS reference station to eliminate systematic bias. Multi-temporal UAV point clouds are co-registered through the Iterative Closest Point (ICP) algorithm to generate Digital Elevation Models (DEMs). An integrated fusion model is then developed by discretizing the monitored area into points and calculating the elevation ratio of each point relative to the GNSS station. Piecewise linear interpolation is used to describe the temporal evolution of these ratios, enabling estimation of surface elevations for any target date based on GNSS time-series data. Results: Field experiments at the Heifangtai landslide (Gansu Province) demonstrated that the unified coordinate framework reduced horizontal deviation between UAV and GNSS results from several meters (maximum 4.69 m) to below 0.2 m. The fusion model successfully generated surface elevation data for dates between UAV flights, with deformation trends showing strong consistency with GNSS time series (Pearson correlation = 0.982). At an independent low-cost GNSS station, the deviation between modeled elevation displacements and actual measurements remained within 1 cm. Deformation rate analysis further revealed a three-phase evolutionary pattern characterized by "acceleration-stabilization-re-acceleration". Conclusions: The proposed UAV-GNSS fusion method effectively overcomes the spatial sparsity of GNSS and the temporal discontinuity of UAV photogrammetry. It enables generation of continuous, centimeter-level landslide surface monitoring data within a unified spatial-temporal framework. This approach provides a practical and reliable solution for high-precision landslide monitoring and early-warning applications.