Abstract: ［Background］Production and construction activities often induce severe soil erosion and ecosystem degradation, especially in ecologically vulnerable regions. Ningxia, located in the upper Yellow River basin in northwestern China, is characterized by pronounced water erosion, low precipitation, and high ecological fragility. The coexistence of intensive construction and environmental sensitivity necessitates systematic evaluation of disturbance and restoration dynamics to support differentiated regulation and targeted ecological restoration. However, existing studies typically address disturbance or restoration in isolation, lacking an integrated framework to quantify the “net” environmental impact across diverse project types.［Methods］This study developed a comprehensive disturbance-restoration evaluation system and applied it to 84 typical production and construction projects completed in Ningxia between 2020 and 2025, covering seven categories: highway, wind power, water diversion, small-scale water conservancy, processing and manufacturing, power transmission and transformation, and real estate. A two-tier indicator system was established, comprising four disturbance indicators and five restoration indicators. The entropy weight method was employed to objectively determine indicator weights, and the Soil and Water Loss Impact Index (SWII) model was extended to calculate separate disturbance and restoration indices, with their sum representing the net impact index.［Results］The entropy weight analysis revealed distinct priorities in evaluating soil erosion processes. Among disturbance factors, earthwork disturbance intensity received the highest weight (0.475), indicating that the volume of excavation and filling per unit area is the most critical driver of initial erosion risk, followed by disturbed land area (0.221). For restoration indicators, the proportion of soil erosion control dominated with a weight of 0.767, demonstrating that achieving the designed soil loss reduction target is the paramount measure of restoration success. The percentage of forestry and grass coverage in the recoverable disturbance zone also played a significant role (0.136), affirming the synergy of engineering and biological measures. Further analysis based on SWII values delineated a clear typological spectrum. Power transmission and transformation projects exhibited the highest disturbance index (0.408) and the highest net SWII (0.335), defining a “high-disturbance-low-restoration” profile that warrants prioritized intervention. Wind power projects (SWII=0.305) and water diversion projects (SWII=0.292) followed, characterized by moderate to high disturbance but insufficient restoration. In contrast, processing and manufacturing projects (0.117) and real estate projects (0.099) achieved the lowest net impacts, demonstrating that restoration efforts within their permissible areas can effectively mitigate disturbance effects. Highway projects (0.180) and other small-scale water conservancy projects (0.209) occupied an intermediate position, suggesting that their current mitigation-restoration balance leaves room for improvement.［Conculsions］The “Entropy Weight Method-SWII” framework developed in this study systematically quantifies the disturbance-restoration dynamics and net soil erosion impacts of production and construction projects in ecologically fragile regions. The findings identify linear infrastructure and energy projects as priority targets for enhanced soil and water conservation supervision and ecological remediation. The model provides a robust quantitative tool to support differentiated project management, restoration prioritization, and soil and water conservation planning. It also offers a transferable methodological reference for evaluating the environmental effects of construction activities in other ecologically vulnerable regions facing similar challenges of balancing development and ecosystem protection.