[Objective] This study attempted to identify and extract anthropogenic disturbance areas resulting from production and construction projects in Hubei Province from 2017 to 2021. Spatiotemporal feature analysis of the classification results was also conducted in order to provide theoretical support and methodological references for the extraction of anthropogenic disturbance areas, and to address soil and water erosion issues. [Methods] The study was conducted in Hubei Province because of its diverse range of anthropogenic disturbance types from production and construction projects. We used the Google Earth Engine (GEE) platform to access Sentinel-2 time-series data. We investigated the optimal combinations of classification features and parameters using a 1 995 dataset. To tackle the issue of differentiating impermeable layers from croplands within anthropogenic disturbance areas caused by production and construction projects, we employed a random forest model for the annual identification and extraction of such areas in Hubei Province from 2017 to 2021. [Results] ① The optimal feature band combination for identifying and extracting anthropogenic disturbance areas from production and construction projects included the red-edge band, green-edge band, blue-edge band, near-infrared band, NDVI, NDWI, NDBI, RRI, dNDVI, contrast, and entropy. ② Overall, the classification accuracy for the years 2017 to 2021 consistently exceeded 93.00%, with kappa coefficients consistently above 0.92, affirming the method was feasibility for extracting anthropogenic disturbance areas due to production and construction projects. ③ The total area of anthropogenic disturbance land parcels in Hubei Province exhibited a pattern characterized as "increase-decrease-increase" from 2017 to 2021, with an anomalous decrease in 2020. [Conclusion] The proposed method demonstrated substantial potential for the rapid identification of large-scale, diverse anthropogenic disturbance areas resulting from production and construction projects. The resulting high-precision, long-term spatial dataset can provide valuable support for subsequent research endeavors related to this topic.