Ocean fronts exert a significant influence on sound propagation in the ocean. We employ a Wentzel-Kramers-Brillouin (WKB) approximation to derive analytical expressions for modal eigenvalues, thereby enabling the evaluation of mode coupling across different fronts. First, idealized models of three distinct fronts are established. Using the WKB method, differences in eigenvalues are calculated to identify regions of quasi-crossing points in eigenvalues, where mode coupling is strongest. The derivation results show that the differences in eigenvalues for vertical fronts decreases monotonically with distance, while those for fronts parallel to the seabed topography either decreases monotonically or remains constant. Only for the third type of front does differences in eigenvalues exhibit a minimum value. Numerical simulations verify the accuracy of the proposed approach. The results indicate that no quasi-crossing points occur when the front is either vertical or parallel to the seabed topography. Moreover, for the third type of front, a quasi-crossing point emerges as a mode transitions from a surface-reflected bottom-reflected mode to a bottom-trapped mode. Numerical simulations confirm that vertical fronts and fronts parallel to the seabed, due to the absence of eigenvalue quasi-crossings, exhibit relatively weak mode coupling. In contrast, for the third type of front, mode coupling intensifies within quasi-crossing regions and remains weak elsewhere. Furthermore, group velocity simulations are performed to elucidate differences in intensity striation patterns among the different fronts. For vertical fronts, although the magnitude of group velocity changes with distance, the group velocity of lower-order modes consistently exceeds that of higher-order modes, producing positive-slope intensity striations. In contrast, fronts parallel to the seabed topography yield negative-slope striations because the magnitude relationship of group velocities reverses over certain distance intervals, with higher-order modes exhibiting greater group velocity than lower-order modes. The intensity striations of the third front are distorted due to intense energy coupling at the quasi-crossing points. Finally, we simulate and explain distinct spatial distributions of acoustic energy for the three different fronts. For vertical fronts, since the sound speed remains constant in the vertical direction, energy is uniformly distributed in the vertical direction. In contrast, for the other two fronts, where the sound speed profile forms a low sound-speed layer near the seabed at the reception range, the energy is trapped beneath the thermocline due to the specific structure of the sound speed profile.