Accurately identifying the spatial distribution of facility agriculture is crucial for agricultural development planning, disaster forecasting, and yield estimation. This study focuses on distinguishing and precisely extracting different types of facility agriculture in Weifang City using the Google Earth Engine (GEE) platform. By integrating Sentinel-1 radar data, Sentinel-2 optical data, and SRTM DEM data, 41 feature variables are extracted, including spectral, index, radar, texture, and topographic features. The texture features are optimized by evaluating 14 different scale windows, with the cumulative difference method identifying the optimal texture features for Sentinel-2 imagery. 6 feature combinations are designed using the Random Forest classification algorithm to generate a 10-meter resolution facility agriculture distribution map for Weifang in 2023, and to explore the impact of different feature combinations on classification accuracy. Results show that: (1) The optimal texture window size for extracting facility agriculture using Sentinel-2 is 27×27, with the best texture feature combination including mean, entropy, variance, dissimilarity, and contrast. (2) When different features are added to spectral and index features, the extraction order of importance is texture, topography, and radar, with the introduction of texture improving classification accuracy by 4.81%. (3) The optimized feature selection yields the highest classification accuracy, with an overall accuracy of 91.03% and a Kappa coefficient of 0.86. The producer"s and user"s accuracy for plastic tunnels are 92.21% and 91.83%, respectively, while those for solar greenhouses are 88.54% and 89.47%, respectively. This research provides a remote sensing extraction method for various facility agriculture types, offering valuable decision-making support for facility agriculture disaster risk management.