Abstract: Based on the analysis of remote sensing images, this paper uses the optimized K-means algorithm to detect environmental changes in Qinghai province, make reasonable plans for the use of land based on the predicted results, finally to achieving sustainable development. This article compares the experimental results with the actual results of field operations and local policy planning. The results show that the algorithm in the article has an accuracy of more than 75% in predicting land changes and future trends, which verifies the feasibility of the algorithm, this ensures that land resources can be used reasonably while the economy is developing, and the environment can be effectively protected.

Key words: change detection, K-means, land use, change prediction, environmental protection