Abstract: In-situ stress affects the permeability of coal reservoirs throughout the extraction of coalbed methane (CBM) and further influences the migration and output of CBM. To explore the effective methods for calculating the in-situ stress in coal seams, this study investigated the southern Yanchuan block at the eastern margin of the Ordos Basin. Based on the logs and well test data of the major coal seams in the study area, as well as the calculation results of parameters including the coal rock Poisson's ratio, modulus of elasticity, and reservoir pressure of coals, this study built a calculation model for the in-situ stress in the coal reservoirs of the study area and analyzed the distribution of in-situ stress in the study area. The results show that: (1) Compared with the combined spring model, the modified Ge’s model is subjected to a lower average relative error of the horizontal principal stress of coal seams and is more applicable to the prediction of the in-situ stress in the southern Yanchuan block; (2) The No.2 coal seam in the study area has a minimum horizontal principal stress ( \sigma _\mathrmh ) of 9.23‒29.36 MPa (average: 16.87 MPa), a maximum horizontal principal stress ( \mathrm\sigma _\mathrmH ) of 10.74‒44.71 MPa (average: 25.49 MPa), and a vertical principal stress (\sigma_ \mathrmv) of 15.64‒39.51 MPa (average: 27.07 MPa); (3) Planarly, the in-situ stress in the coal reservoirs of the study area is high in the northwest and low in the southeast with the Xizhang fault zone as the boundary. Vertically, with an increase in the burial depth of coal seams, \sigma _ \mathrmv, \sigma _ \mathrmH, \sigma _ \mathrmh, and the lateral pressure coefficient all show an upward trend, and the stress field state gradually transforms from the normal fault type ( \sigma _\mathrmv > \sigma _\mathrmH > \sigma _\mathrmh ) to the strike-slip type ( \sigma _\mathrmH ≈ \sigma _\mathrmv > \sigma _\mathrmh ). Moreover, the changes in the lateral pressure coefficient and stress field state of the coal reservoirs in the study area are closely related to the distribution of small reverse faults. The modified Ge's model has a prediction accuracy of 92.15% and can be applied to other blocks by modifying the tectonic stress coefficients. Therefore, this model holds great significance for the extraction of deep CBM in the study area and even other blocks.