Abstract: Aiming at the problem that the index weights in the existing aquifer water-richness evaluation are all constant weights,and the changes between single main control factors and the changes under the combination of multiple factors are ignored. Under the guidance of “water-richness index method” idea,a water-richness index method based on partitioned variable weight theory is proposed,which can not only determine the different evaluation indicators weight,but also can determine different weights when the same indicator is in different state values. Based on the lithology differences,hydraulic characteristics,structural factors and geophysical parameters that affect the water richness of the aquifer,a system of main control factors of 4 categories and 16 sub-categories is established that can more fully describe the distribution law of water-richness. The method of standardization and dimensionless treatment of the index values of the main control factors are summarized,which lays a foundation for multi-factor information fusion analysis. The authors use variable weights theory and K-means clustering algorithm of dynamic clustering to divide variable weight range threshold of each index of main factors,and construct a conclusive theoretical method of state variable weight vector and adjustable weight parameters. The interaction and relative importance of each main control factor to the water-rich aquifer are determined quantitatively. The evaluation method and model of water-rich index based on variable weight theory is constructed. At the same time,taking the sandstone fissure aquifer of Xiaozhuang coal mine in Binchang mining area as an example,the specific implementation steps of water-rich index method based on zoning and variable weights are explained. Five sets of pumping test data are used to correct the evaluation results. The results show that the water-rich aquifer of the Luohe Formation water-filled aquifer in Xiaozhuang coal mine demonstrates a decreasing trend from northwest to southeast. Stronger-strong water-rich areas are mainly distributed in the second and third panels,and the western boundary of the first basin is also distributed in a small area. The middle rich water area is distributed in the west of the central part of the mining area and the northwest of five panels in a strip. The rest areas are weak to weak rich water area. Finally, compared with the traditional constant weight evaluation results,which proves that the evaluation results of variable weight can better reflect the characteristics of aquifer heterogeneity and the prediction accuracy is higher.