Abstract: Deep coalbed methane (CBM) production behavior differs markedly from that of shallow CBM. Coal seam burial depth, as a core factor influencing reservoir physical properties, fluid occurrence, and seepage patterns, plays a particularly important role in regulating gas and water production. In-depth analysis of the factors influencing the differences in production characteristics between deep and shallow CBM, as well as the depth effect on production dynamics, is of great significance for productivity prediction and production in subsequent CBM development stages. To clarify the depth effect on the dynamics of deep and shallow CBM in the Daning-Jixian Block, research was conducted based on geological and engineering factors such as coal seam burial depth, gas content, reservoir pressure, vitrinite reflectance, structural characteristics, hydrochemical characteristics, and fracturing stimulation intensity. The results show that compared to shallow coal reservoirs, deep coal reservoirs in the Daning—Jixian Block are characterized by “high gas content, high adsorption saturation, high reservoir pressure, and low permeability,” and there are significant differences in the gas and water production dynamics between deep and shallow CBM wells. Using a support vector machine (SVM) regression model and a data augmentation strategy, a CBM productivity prediction model integrating multiple geological and engineering factors was constructed, achieving high-precision prediction of stable gas production rates in both shallow and deep CBM wells. The model achieved an R² of 0.91 on the test set for shallow samples and 0.90 for deep samples, demonstrating good fitting capability and generalization performance. Based on comprehensive analysis of multiple geological and engineering parameters and machine learning SHAP attribution analysis, the main controlling factors of production characteristics in deep and shallow CBM wells and their contribution sequences were identified. Shallow CBM productivity is mainly dominated by engineering stimulation intensity, with the contribution of main controlling factors in the following order: total fracturing scale, salinity, fracture intensity coefficient, gas content, burial depth, reservoir pressure, and vitrinite reflectance. This indicates that in shallow areas, artificial fracturing stimulation is the core means to break through reservoir physical limitations and enhance productivity. In contrast, deep CBM productivity is dominated by reservoir geology and fluid properties, with the contribution of main controlling factors in the following order: salinity, total fracturing scale, gas content, burial depth, reservoir pressure, vitrinite reflectance and fracture intensity coefficient. This highlights the key controlling role of reservoir fluid properties and deep enclosed hydrogeological conditions in maintaining high and stable production. The above research results reveal the depth effect and main controlling factors of CBM production dynamics in the Daning—Jixian Block, providing reference significance for CBM production in the study area and adjacent regions.