In the thesis the use and accuracy of reverse engineering in the mechanical woodworking process were examined. For acquiring the shape of existing objects photographing was used and the captured images were processed in the computer software Agisoft Photoscan, where photos were aligned in space and a 3D model created. Using SolidWorks and SolidCAM a 3D model was imported and a G-code for CNC machine was then generated. Oak, elm and maple wood were used to produce the replicas, where elm and maple proved to be the most suitable, while oak, due to its slightly coarser structure, produced worse surface quality. The quality of the finished surface of the product is most affected by the light, the surface of the photographed object, and the choice of the alignment accuracy of the photos in the software itself. The light should be as uniform as possible and sufficient. When photographing the surface of the object, it is preferable that the light reflecting off said object is diffused and not reflected. It is advisable to choose a high degree of precision in the alignment stages of the program, as the resulting surface quality is improved. The reverse engineering method is useful in woodworking because it is more time efficient than different methods of measurement and 3D modeling.