Associated with reservoir monitoring towards optimizing production of oil and gas and maximizing the recovery, or ensuring safe CO2 storage, it may be beneficial to introduce integrated geomechanics, rock physics and seismic modelling in the work flow. The geomechanics modelling estimates the effective stress change and deformation in the reservoir and surrounding rocks induced by reservoir pore pressure change. The rock physics modelling estimates the subsequent change of the seismic wave velocities in the rock. Then the seismic modelling estimates detectable changes in the seismic signals. Rock anisotropy may affect every component in such integrated modelling. In this paper we performed the integrated modelling study and compared two simplified cases: In one case, all the rocks were assumed isotropic in stiffness, while in another case, the overburden and sideburden were assumed transeversely isotropic. The geomechanical modelling showed that the in situ stresses which were assumed pure results of rock gravity were already different in two cases. After reservoir depletion, the stress changes were different in two cases as well. A rock physics model was used to calculate the changes of seismic wave velocities as results of the changes of rock stresses and rock deformation. The rock physics model is based on the diffenrential effective medium theory and assumes the rock as a solid with spherical void inclusions and penny-shaped cracks, aiming at capturing stress-sensitive and anisotropic wave velocities. Seismic wave two way travel times and time shifts were calculated along different paths, with different offsets but the same reflection midpoint. The results show the difference between the isotropic case and the anisotropic case. Such difference increases with increasing offset. We conclude that such integrated modelling is important to understand the physical processes in the subsurface because those processes are interdependent. Rock anisotropy plays an important role in the physical processes and has significant impacts on the results of the integrated modelling.