This thesis presents a set of methods to enhance the automation on 3D architectural and structural modeling for historic timber structures using terrestrial laser scanning (TLS) point clouds. The preservation of historic timber constructions, which serve as invaluable cultural heritage assets, requires effective structural health monitoring (SHM) to ensure their safety and longevity. SHM relies on detailed geometric data acquisition and 3D modeling to detect potential structural deviations. However, the manual modeling of each structural element in computer-aided design (CAD)-based structural analysis software is a labor-intensive and error-prone process, whereas fully automated modeling often yields incomplete results due to challenges in data acquisition and processing.Fitting cuboids onto the point clouds of straight beams with rectangular cross-sections is employed in order to fill the gap between point clouds and architectural and structural models. Building on this, an advanced workflow incorporating a linear sub-segment split method was developed, with the result that the completeness of automated modeling was significantly enhanced. Additionally, a novel approach was implemented to detect and remove non-structural roof cover points. This workflow was applied to an entire historic timber roof structure, demonstrating a notable improvement in quantitative completeness.Subsequently, the thesis focuses on identifying planar roof tiles and sub-structural systems while enhancing model integrity through operations such as creation, merging, and extension using a logic-based decision mechanism. The resulting models were successfully integrated into structural analysis software, enabling load case simulations and facilitating preliminary structural assessments. A comparative analysis with manually generated models in RSTAB and Revit, highlights both the advantages and limitations of the proposed automated approach in contrast to traditional modeling methods. Furthermore, the beam modeling technique is expanded by incorporating trapezoidal prisms in addition to cuboids, further refining the accuracy and versatility of the automated modeling process.The results show that the proposed automated workflows significantly reduce the time required for preliminary structural assessments, decreasing the need for labor-intensive work while also minimizing errors resulting from human intervention. The automation level in the state-of-the-art approach has been greatly improved, and the concept has been successfully validated on different structures. This thesis highlights the efficiency of automated 3D modeling in early-stage structural evaluations, providing a practical and scalable approach for the documentation, analysis, and preservation of historic timber structures.