Abstract: The exploitation and utilization of mineral resources play a very important role in supporting the economy of the less developed inland provinces and regions, and the rational exploitation and utilization of mineral resources become the key way of sustainable development. Mineral resources are widely distributed in Ningxia Hui Autonomous Region, but illegal exploitation, represented by cross-border mining, restricts the development of local economy. Traditional methods of inspection, reporting and public reporting make it difficult to obtain information about large-scale cross-border mining, while remote sensing data is considered to be an effective supervisory method for monitoring cross-border mining behavior due to its extensiveness and real-time performance. Based on the relevant information and the GF2 image, this paper adopts object-oriented method to segment, extracts the imaging spectrum and spatial texture statistical features of the plot, and uses classification and regression tree, random forest and support vector machine and other machine learning classification algorithms to identify the cross-border mining surface of open-pit minerals. The statistical results show that the total accuracy of the three classifiers is above 60%, and the recognition effect of random forest classifier is the best, which can reach 75%. The verification results show that there are overclassification cases in the identification of cross-border mining, and the target area should be further screened based on mining rights data in practical application. This paper uses remote sensing satellite images to identify and study the cross-border mining surface, which greatly improves the efficiency of mine law enforcement and promotes the sustainable development and utilization of mineral resources and mine environmental protection.