In recent years, 3D mesh models have gained significant importance in various applications, including computer graphics, digital watermarking, and secure communications. However, ensuring high embedding capacity, robustness, and security in blind data hiding remains a challenging task. This paper proposes an efficient blind data hiding approach based on digit substitution for 3D mesh models, designed to enhance both security and robustness against geometric attacks. The proposed method begins by extracting the vertices from the 3D mesh model, followed by the application of preprocessing to achieve transformation invariance. To further enhance security, breadth-first search (BFS) is employed to generate a unique sequence, ensuring a secure embedding process. A novel digit substitution technique is then introduced to determine reference positions suitable for embedding secret data effectively. To increase the embedding capacity, secret data is transformed into a octal representation by grouping every three-bit set before embedding. Extensive experimental analysis demonstrates that the proposed approach achieves a high embedding capacity of 9.6 bpv while maintaining superior imperceptibility, as reflected in high Peak Signal-to-Noise Ratio (PSNR) values. Furthermore, the technique exhibits strong resilience against various attacks, including geometric transformations.