In the field of computer vision, single-view 3D object reconstruction faces numerous challenges. Traditional approaches often rely on multi-view images or additional depth information, limiting their effectiveness and flexibility in real-world applications. Especially in cases where depth information is absent, the accuracy and robustness of reconstruction results are frequently compromised. Such challenges prompt researchers to develop more efficient and accurate methods to tackle complex scenarios involving varying object shapes, view dependencies, and occlusions. To address these issues, this paper proposes a single-view 3D object reconstruction method based on category and part priors. The approach first employs a convolutional neural network (CNN) to extract deep features from a single RGB image. Then, category and part priors are introduced to guide the model in reconstructing object shapes and structures. Concurrently, textual descriptions associated with the image are encoded using a text processing model to supplement visual information. Subsequently, the shape reconstruction module transforms these fused features into a 3D model, progressively generating high-resolution 3D shapes through deconvolution and upsampling techniques. Finally, shape optimization algorithms further refine the reconstructed result, ensuring that the outputted 3D object is both accurate and representative of the real object’s shape.