Abstract: The classical highlight model method is primarily applicable for the target strength (TS) computation of simple targets. To extend its applicability to predicting scattering characteristics for complex targets, a sliced highlight model method was previously proposed. However, the accuracy of the highlight model method remains insufficient compared to the Kirchhoff approximation. To overcome this limitation, this study proposes a modified highlight model method that introduces equivalent reflection coefficients for different components of underwater vehicles and utilizes a genetic algorithm (GA) to invert the surface reflection coefficients of each component. This approach significantly enhances the prediction accuracy of complex target acoustic scattering characteristics while preserving the computational efficiency advantage of the highlight model. The proposed method yields an average error below 4 dB compared to the planar element method, with a computational efficiency improvement of about 1000 times, offering a practical solution for real-time prediction of underwater vehicle acoustic scattering characteristics.