Compared with the Palm line structure features, extraction and description of palm print texture features are easier. But, with the increase in the number of palmprint samples, these features are not powerful enough. In order to solve the problem, the paper proposes a new approach to enhance the distinguishing capability of texture features for palm print recognition. It uses classical results on Riemannian geometry to obtain the information of palm lines and construct direction fields of palm lines. The direction fields become a part of the textures of the palmprint image to enhance the distinguishing capability of texture features. Finally, the dual-tree complex wavelet transform-based local binary pattern weighted histogram method (DT -CWT based LBPWH) is used to extract enhanced texture features. The experimental results validate the effectiveness of the method.