Coal seams with a high CO2 content may have outburst risk, and degasification of CO2 has to be conducted before these coal seams can be safely extracted. For geosequestration of CO2 in unminable coal seams, injected CO2 may desorb with the reduction of CO2 pressure. Desorption of CO2 dominates these processes, while adsorption isotherms are widely used assuming that the adsorptiondesorption process is fully reversible. To understand the difference between CO2 adsorption and desorption isotherms, i.e., sorption hysteresis, as well as the dependence of CO2 sorption hysteresis on maximum pressure, four cycles of CO2 adsorptiondesorption experiments are conducted continuously with increasing maximum pressure (1, 2, 3, and 4 MPa). The difference of CO2 emission volume between adsorption and desorption isotherms is compared, and a significant deviation (0.059 and 0.032 mol/g) has been observed. The adsorption isotherms show a good repeatability, indicating that the gas holding capacity does not change during a long-term contact with CO2. However, a distinct difference between the desorption isotherms is observed. The path of desorption isotherm depends upon the maximum pressure, and higher maximum pressures can reduce the proneness of CO2 desorption in the pressure range of this study (04 MPa). We suggest that desorption isotherms should be used to predict the CO2 emission volume and long-term storage stability, and the maximum pressure of the laboratory sorption test should be decided according to the in situ coal seam pressure.