Seismic waves are used for the detection of reservoirs and material boundaries in presence of fracture. Finite difference method (FDM) model the elastic waves in homogenous medium. The physical boundary is set to rigid boundary condition while the computational boundaries are computed on Clayton Engquist (CE) boundary condition. After getting the synthetic gathers, a preprocessing step i.e. automatic gain control (AGC) is used to rise the subsurface reflections. Then autoregressive of order one is employed to model the subsurface reflections. The estimated autoregressive model for synthetic data explains the 94.739 variability and amplitude value of -5.5798e-09 similar to real data -1.09173e-09 when all the values in predictor are zero. The comparison with real model provides evidence that predicted model can interpret the acquisition geometry material properties and its boundary in the real subsurface. © 2016 Umairia Malik et al.