The city of Qinzhou, serving as the primary construction site for the century project "Pinglu Yunhe", exploits modern remote sensing technology to generate high-precision land use/cover change (LUCC) datasets, which exert substantial implications for its national spatial planning and ecological protection and construction. This research, concentrating on the Beibu Gulf coastal area and taking Qinzhou as an exemplar, employs Google Earth Engine (GEE) and Landsat remote sensing images to construct a LUCC dataset during the period from 2012 to 2022 and analyze its temporal and spatial pattern evolution by means of a random forest model integrating spectral, texture, index, and terrain features. The study also introduces the optimal parameter geographic detector to explore the driving mechanisms. The results demonstrate that: (1) The optimized random forest model can effectively extract remote sensing information, with overall accuracies (OA) ranging from 0.88 to 0.92 and Kappa coefficients ranging from 0.86 to 0.90. The overall accuracy and Kappa coefficient of each period"s LUCC product are in line with the classification results of several 4km×4km interpretation patches and Google Earth"s high-resolution images of the same period and location, indicating that the classification results of land features are consistent with the actual land features. (2) From 2012 to 2022, the area of forest land in Qinzhou increased by 91.93km2, the area of cultivated land decreased by 284.73km2, and the area of built-up land rose by 180.05km2. The overall trend of land use and cover change is ascending. (3) During the study period, the land use and cover change in Qinzhou was mainly driven by economic factors (GDP) and terrain features (DEM and slope), with the two factors presenting a dual-factor enhancement effect on the temporal and spatial pattern evolution in the three phases of 2012-2017, 2017-2022, and 2012-2022 respectively. The interactive effect between ecological factors (NPP and precipitation) and GDP shows different degrees of dual-factor enhancement at different stages. This study can provide data support for land resource management in Qinzhou and the produced LUCC datasets can serve as a reference for guidance.