Industrial computed tomography (CT) has emerged as a critical non-destructive testing tool for inspecting large-scale components in challenging industrial environments, particularly in nuclear engineering. However, the inherent complexity in size, shape, and internal structure often renders conventional full-angle scanning impractical, resulting in incomplete projection data and severe imaging artifacts that compromise defect detection and structural integrity assessment. In response to these challenges, this study proposes an integrated approach that combines optimized limited-angle scanning trajectories with a novel reconstruction algorithm—Contour Guided-Deep Radon Prior (CG-DRP). Utilizing numerical simulations on a 1:1 scale wooden Dougong model (with an assumed homogeneous density of 0.5 g/cm3), three representative limited-angle scanning trajectories were designed to maximize projection coverage while conforming to practical operational constraints. The proposed CG-DRP algorithm, which enhances the unsupervised Deep Radon Prior (DRP) framework by incorporating geometric contour constraints, effectively compensates for missing projections and suppresses artifacts, thereby significantly improving reconstruction accuracy and robustness under noisy conditions. Comparative evaluations against traditional reconstruction methods—including Filtered Back Projection (FBP), Simultaneous Algebraic Reconstruction Technique (SART), and Alternating Direction Method of Multipliers-Total Variation (ADMM-TV)—demonstrate the superior performance of CG-DRP algorithm. These findings underscore the potential of the proposed approach to advance industrial CT inspection capabilities for large, complex components, particularly in scenarios where complete angular coverage is unattainable, and lay a solid foundation for future research in adaptive trajectory optimization and algorithm enhancement in limited-angle CT imaging.