We present a novel scheme for high-capacity of 3-D triangle meshes. Our proposal can be applied to meshes with arbi- trary topology by using irregular wavelet-based analysis. The secret message is inserted in an appropriate resolution level by quantizing the norms of wavelet coefficients vectors. In addition, a re-ordering process is applied as pre-processing in both embedding and extraction stages. Simulation results show that our data hiding framework is robust to common geometric attacks including the operations of the similarity transformations and achieves high data capacity with insignif- icant visual distortion.