The main methods of coalbed methane (CBM) development are drainage and depressurization, and a precise prediction of coal reservoir pressure is thus crucial for the evaluation of reservoir potentials and the formulation of reasonable development plans. This work established a new reservoir pressure prediction model based on the material balance equation (MBE) of coal reservoir, which considers the self-regulating effects of coal reservoirs and the dynamic change of equivalent drainage area (EDA). According to the proposed model, the reservoir pressure can be predicted based on reservoir condition data and the actual production data of a single well. Compared with traditional reservoir pressure prediction models which regard EDA as a fixed value, the proposed model can better predict the average pressure of reservoirs. Moreover, orthogonal experiments were designed to evaluate the sensitivity of reservoir parameters on the reservoir pressure prediction results of this proposed model. The results show that the saturation of irreducible water is the most sensitive parameter, followed by Langmuir volume and reservoir porosity, and Langmuir pressure is the least sensitive parameter. In addition, the pressure drop of reservoirs is negatively correlated with the saturation of irreducible water and the Langmuir volume, while it is positively correlated with porosity. This work analyzed the reservoir pressure drop characteristics of the CBM wells in the Shizhuangnan Block of the Qinshui Basin, and the results show that the CBM reservoir depressurization can be divided into three types, i.e., rapidly drop type, medium-term stability type, and slowly drop type. The drainage features of wells were reasonably interpreted based on the comprehensive analysis of the reservoir depressurization type; the latter was coupled to the corresponding permeability dynamic change characteristics, eventually proving the applicability of the proposed model.