Abstract: The complex mineral distribution, particularly clay minerals impacting reservoir sensitivity, poses a key challenge for optimizing development and preventing damage in the Junggar Basin's coalbed methane reservoirs. To address this, coal samples from four wells were analyzed using full-diameter CT scanning and field emission SEM. Shape factor and CT grayscale value were introduced as key parameters, establishing a method to distinguish coal matrix, clay minerals, pyrite, and calcite. This method was applied for mineral modeling, content calculation, sensitivity evaluation, and gas content prediction. Results show: (1) Combined CT grayscale value and shape factor analysis effectively distinguishes minerals; calcite’s high shape factor (10-50) is a key identifier. (2) A reservoir sensitivity classification based on clay content was developed: weak (＜40%), medium (40%-70%), and strong (＞70%) water sensitivity. (3) A significant negative correlation model (R＞0.7) between clay content and measured gas content was built, with prediction errors ＜10%. (4) Controlled depressurization rates are recommended for medium-strong sensitivity zones during drainage. Key breakthroughs include revealing vertical heterogeneity and clay enrichment via continuous quantification, coupling spatial clay distribution with sensitivity for drainage control guidance, and supporting “sweet spot” identification via the gas content prediction model.