Rocky desertification is a core issue restricting the coordinated development of the ecological environment and socio−economy in the karst regions of southwestern China. To address the limitations of traditional evaluation methods—such as high subjectivity and difficulty in quantification—this study proposes an intelligent quantitative evaluation model for rocky desertification by integrating logistic regression and spatial analysis techniques. A case study was conducted in Bijie City, Guizhou Province. Based on the evolution process and formation mechanism of rocky desertification, this study selects 12 candidate factors from both natural and human dimensions. Logistic regression models are then used to analyze the significance of these factors' influence on rocky desertification and the threshold for classifying its states. Ultimately, five key factors—vegetation cover, bedrock exposure, slope, surface roughness, and profile curvature—are identified as the foundation for model construction. The evaluation model integrates the entropy weight method to adaptively assign weights to the factors, and intelligently accomplishes the spatial distribution and severity classification of rocky desertification on the GIS platform. Accuracy validation shows a classification accuracy of 82.6% and a Kappa coefficient of 0.76, effectively revealing the spatial distribution and differentiation patterns of rocky desertification in Bijie. This method, through threshold quantification, weight optimization, and multi−source data fusion, provides foundational data for dynamic monitoring and differentiated management of rocky desertification in karst areas. It holds significant practical value for advancing ecological restoration and regional sustainable development.